JP4317032B2 - Vehicle rollover detection system - Google Patents

Description

Detailed Description of the Invention

When a vehicle rollover occurs, especially in a type of rollover with a relatively short head closure time, before the head first contacts the vehicle interior, for example, a seat belt pretensioner, air There is a need for a vehicle rollover detection system that allows vehicle rollover discrimination with time to deploy an associated safety restraint actuator such as a bag or roll curtain. For example, depending on the rollover event, a head contact may occur before it can be reliably determined from the physical laws of the roll rotation event whether the vehicle will roll over completely. There is also a need for a robust vehicle rollover detection system that determines vehicle rollover sufficiently quickly depending on whether the rollover is a relatively slow or relatively fast event.
Referring to FIG. 1 a, a rollover detection system 10 is mounted on a vehicle 12. The vehicle 12 is shown in a local Cartesian coordinate system, where the X axis coincides with the longitudinal axis of the vehicle with the front as positive, the Y axis with the lateral axis of the vehicle with the left direction as positive, and the Z axis as It coincides with the vertical axis of the vehicle with the upward direction being positive. When the mass of the vehicle 12 is M, the center of gravity CG corresponding to it is located at a height of Z ₀ from the ground. In the figure, the vehicle 12 slides at a speed U toward the obstacle 14 in the negative Y-axis direction.

が、係合以前の位置と方向に対して増大する。つまり、車両１２の位置エネルギは、その角度位置θに依存する。また、回転と共に車両１２には、回転エネルギ

が生じる。反力Ｆによっても、Ｙ軸に沿った横加速度成分Ａ_ｙ（ｔ）で示すように、重心ＣＧの直線加速度

が生じる。図１ａおよび１ｂは、車両がスライドして障害物と衝突することによって生じるロール事象を示すが、例えばタイヤの破裂によって、その後に関連する車輪リムが接地するなど、ロール事象は別の状況によっても生じる可能性があることはいうまでもない。したがって、ロールオーバ検出システム１０は、特定の種類のロール事象に限定されるものではない。 Referring to FIG. 1 b, when one or more wheels 16 of the vehicle 12 are engaged with the obstacle 14, the reaction force F generated from the vehicle 12 causes the vehicle 12 to change over time about the X axis about the trip point 13. Rotating at an angular velocity ω _x (t) to produce an angular position θ (t) that changes over time. Here, the inertia moment of the vehicle 12 around the rotation axis concerned is I _x , and this rotation axis is parallel to the X axis and crosses the trip point 13. Due to the rotation of the vehicle 12, the height Z _CG of the center of gravity CG increases with respect to the previous height Z ₀ of contact with the obstacle 14, whereby the potential energy of the vehicle 12 is increased.

Increases with respect to the position and direction prior to engagement. That is, the potential energy of the vehicle 12 depends on the angular position θ. In addition, the rotational energy is transmitted to the vehicle 12 along with the rotation.

Occurs. Even with the reaction force F, as indicated by the lateral acceleration component A _y (t) along the Y-axis, the linear acceleration of the center of gravity CG

Occurs. FIGS. 1a and 1b show a roll event caused by a vehicle sliding and colliding with an obstacle, but the roll event can also be caused by other circumstances, for example, a tire rupture and then the associated wheel rim touches down. Needless to say, this can happen. Accordingly, the rollover detection system 10 is not limited to a particular type of roll event.

Referring to FIG. 2, the rollover detection system 10 includes a lateral accelerometer 18 and an angular velocity sensor 20, which are not required, but are preferably mounted close to the center of gravity CG of the vehicle 12. The lateral accelerometer 18 is responsive to a time-varying lateral acceleration component A _y (t) along the local Y axis. For example, the lateral accelerometer 18 may be an accelerometer having one sensitivity axis that substantially matches the local Y axis, such as a micromachined accelerometer having at least one axis sensitivity. The orientation of the angular velocity sensor 20, such as a gyroscope, is determined so as to be responsive to the time-varying component of the angular velocity around the local X axis. Lateral accelerometer 18 and angular velocity sensor 20 are operatively connected to respective filters 22, 24 for processing respective signals A _y (t) and ω _x (t) by processor 26 having memory 28. Filter with these filters. These filters 22, 24 may be separate from the processor 26, embedded, or analog, digital, or combinations thereof, as is known to those skilled in the art. Should be understood. The filters 22 and 24 may be adapted to be part of the respective lateral accelerometer 18 or angular velocity sensor 20.

を処理し、車両がロールオーバする可能性があるかどうかを判別し、その結果に応じて、適切な安全拘束アクチュエータ３０の作動を制御して、車両１２の乗員へのロールオーバ傷害を軽減する。プロセッサ２６は例えば、デジタルコンピュータ、マイクロプロセッサもしくはその他のプログラム可能なデバイス、アナログプロセッサ、アナログもしくはデジタル回路、またはそれらの組合せとすることができる。さらに、安全拘束アクチュエータ３０としては、それに限定されるわけではないが、シートベルト３４に動作可能に接続されたシートベルトプリテンショナ３２、ロールオーバおよび側面衝突事故の両方からの保護をする胸部エアバッグインフレータ３６、乗員と車両１２の側面ウインド３９との間に展開するように適合されたロールカーテン３８、車両１２のルーフまたはヘッドライナからエアバッグを展開するように適合させたオーバヘッド型エアバッグインフレータ４０などを含めてもよい。図２では車両１２の一方の座席についての安全拘束アクチュエータを示してあるが、安全拘束アクチュエータはすべての座席に備えることが可能であり、関連する安全拘束アクチュエータ３０が、乗員の傷害を軽減するように適合されている、すべての方向のロールオーバに応じて、これらの安全拘束アクチュエータ３０の任意の１つ、あるいはすべてを、ロールオーバ検出システム１０で制御するように適合させることができる。また、特定の安全拘束アクチュエータ３０の組合せが、上記のすべての安全拘束アクチュエータを必ずしも含む必要はなく、上記以外の他形式の安全拘束アクチュエータ３０を含んでもよい。 The processor 26 detects each filtered signal

To determine whether the vehicle is likely to roll over, and to control the operation of the appropriate safety restraint actuator 30 according to the result, to reduce the rollover injury to the occupant of the vehicle 12 . The processor 26 can be, for example, a digital computer, a microprocessor or other programmable device, an analog processor, an analog or digital circuit, or a combination thereof. Further, the safety restraint actuator 30 includes, but is not limited to, a seat belt pretensioner 32 operably connected to the seat belt 34, a chest airbag that protects against both rollover and side collisions. An inflator 36, a roll curtain 38 adapted to deploy between an occupant and a side window 39 of the vehicle 12, an overhead airbag inflator 40 adapted to deploy an airbag from the roof or headliner of the vehicle 12. Etc. may be included. Although the safety restraint actuator for one seat of the vehicle 12 is shown in FIG. 2, the safety restraint actuator can be provided in all seats so that the associated safety restraint actuator 30 reduces occupant injury. Any one or all of these safety restraint actuators 30 can be adapted to be controlled by the rollover detection system 10 depending on the rollover in all directions. Further, the combination of the specific safety constraint actuators 30 does not necessarily include all the above-described safety constraint actuators, and may include other types of safety constraint actuators 30 other than those described above.

信号およびフィルタリングされた

信号の両方を使用して、別の基準と共に展開決定を行うために用いる閾値と比較するための関数が評価される。 Referring to FIG. 3, an embodiment of a rollover detection algorithm 100 that detects rollover of a vehicle and controls the operation of one or more associated safety restraint actuators 30 by, for example, the apparatus shown in FIG. Are signals that control the operation of the data collection and processing algorithm 150, measures algorithm 300.1, energy algorithm 300.2, safety algorithm 200, and safety restraint actuator (s) 30 accordingly. A combination of associated logic 330 ′, 340 that generates 342 is included.
The measure algorithm 300.1 uses an empirical time domain discrimination process to detect the rollover condition, usually with a large lateral force of the vehicle, and a relatively fast head contact time ( For most rollover events characterized by <250 msec), for example, it is advantageous in reducing deployment time. In measure algorithm 300.1, filtered

Signal and filtered

Both signals are used to evaluate a function for comparison with a threshold used to make an expansion decision with another criterion.

信号と、フィルタリングされた

信号の両方を使用する。 The energy algorithm 300.2 detects a rollover state using a phase space discriminating process based on a physical law relating to a vehicle rollover occurrence process, and is mainly low in the vehicle vertical force or the vehicle 12. This is advantageous for making reliable deployment decisions for slow roll events caused by lateral forces. The energy algorithm 300.2 utilizes the filtered angular velocity signal to identify the roll state of the vehicle 12, and to exceed the instantaneous total energy (rotational kinetic energy and potential energy) and the associated equilibrium point. The energy required for rotating the roll is compared. The energy algorithm 300.2 was filtered to the relevant entry and exit criteria

Signal and filtered

Use both signals.

およびフィルタリングされた

のデータを利用する。測度アルゴリズム３００．１およびエネルギアルゴリズム３００．２の両方とも、関連する入口基準および出口基準に特徴があり、それぞれのアルゴリズムに関連する計算は、それぞれの関連する入口基準が満たされた場合に開始され、それぞれの関連する出口基準が満たされた場合には、これらの計算は停止されて、その後に入口基準が満たされるとリセットされる。 Although FIG. 3 shows a combination of the measure algorithm 300.1 and the energy algorithm 300.2, this is not essential and it is understood that either algorithm can be used alone. Should. However, combining the algorithms can increase the robustness of the associated rollover detection system 10. This is because, for example, in a state such as a “curb-trip” state, the measure algorithm 300.1 can perform a more rapid discrimination than the energy algorithm 300.2, but in contrast to this, the “spiral” In other states, such as “corkscrew”, “ramp”, or “flip”, the energy algorithm 300.2 provides faster discrimination than the measure algorithm 300.1. Is possible.
The measure algorithm 300.1 and the energy algorithm 300.2 are independent of each other, but both are filtered common data from the data collection and processing algorithm 150, ie filtered

And filtered

Use the data. Both measure algorithm 300.1 and energy algorithm 300.2 are characterized by associated entry and exit criteria, and the computation associated with each algorithm is initiated when each associated entry criterion is met. If each associated exit criterion is met, these calculations are stopped and then reset when the entrance criterion is met.

および／またはフィルタリングされた

に依存する、独立した条件セット、または安全化基準を提供ことによって、ロールオーバ検出システム１０の信頼性を向上させることができる。測度アルゴリズム３００．１とエネルギアルゴリズム３００．２の両方とも、共通の安全化アルゴリズム２００によって、それぞれ「安全化」される。安全化アルゴリズム２００によって、追加の判別を行い、非ロールオーバ事象に対しての安全拘束アクチュエータ３０の不要な作動を緩和することが可能となるが、安全化アルゴリズム２００は本質的なものではなく、測度アルゴリズム３００．１およびエネルギアルゴリズム３００．２は、いずれも単独で使用することも、互いに組み合わせて使用することも可能であり、また安全化アルゴリズム２００を伴っても、伴わなくても使用可能であることを理解すべきである。 The safing algorithm 200 is a filtered, which must be satisfied in order for one or more corresponding safeguarded actuators 30 to be deployable.

And / or filtered

By providing an independent set of conditions, or safety criteria, that depend on, the reliability of the rollover detection system 10 can be improved. Both the measure algorithm 300.1 and the energy algorithm 300.2 are each “secured” by the common safety algorithm 200. Although the safety algorithm 200 makes additional determinations and can mitigate unnecessary actuation of the safety restraint actuator 30 for non-rollover events, the safety algorithm 200 is not essential, Both the measure algorithm 300.1 and the energy algorithm 300.2 can be used alone or in combination with each other, and can be used with or without the safety algorithm 200. It should be understood.

In operation of rollover detection algorithm 100, depending on the data from data collection and processing algorithm 150, measure algorithm 300.1 or (OR 330 ′) energy algorithm 300.2 detects the rollover condition of the vehicle and (AND 340). ) If the safety algorithm 200 determines that the associated independent safety condition is met, in step (350), the relevant association that may occur due to the rollover event, whether or not the vehicle 12 actually rolls over. One or more safety restraint actuators 30 are deployed to reduce injury to vehicle occupants.
The data collection and processing algorithm 150, the security algorithm 200, the measure algorithm 300.1, and the energy algorithm 300.2 will be described below with reference to the flowcharts shown in FIGS. FIG. 6 shows a flowchart of the general algorithm structure of both measure algorithm 300.1 and energy algorithm 300.2, with individual details of measure algorithm 300.1 and energy algorithm 300.2 shown in FIGS. Is shown in the table. The algorithm is described mathematically and uses parameters for constants specific to the application, and these parameters are shown in FIGS. 9a and 9b, along with example values for a particular type of vehicle. is there. These parameters are usually adapted to a particular application, for example a vehicle platform, and the specific numerical values of the parameters shown in FIGS. 9a and 9b are for illustration only and are within the scope of the present invention. It goes without saying that it should not be considered as limiting.

であり、それに伴う横加速度成分Ａ_y(t)は一般に±２０ｇ

の範囲にあることがわかった。これらのそれぞれの限界値を超える横加速度成分Ａ_ｙ(t)および角速度ω_ｘは、それぞれステップ（１５４）および（１６０）において、それぞれその値で刈り込まれる。例えば、関連する範囲が±２０ｇの例については、−２０ｇ未満の横加速度成分Ａ_ｙ(t)の測定値は、ステップ（１５４）においてその値が−２０ｇに設定される。横加速度計１８と角速度センサ２０の極性は、角速度ω_ｘ信号および横加速度成分Ａ_ｙ信号の対応する極性が、ロール発生中は互いに同じになるように設定される。一般に横加速度計１８からの信号を刈り込むレベル

は、２０ｇまたは横加速度計１８の測定範囲のいずれかの最小値に設定される。同様に、角速度センサ２０からの信号を刈り込むレベル

は、３００度／秒または角速度センサ２０の測定範囲のいずれかの最小値に設定される。 Referring to FIG. 4, the data collection and processing algorithm 150 obtains the measured value of the lateral acceleration component A _y from the lateral accelerometer 18 in step (152), and the longitudinal angular velocity ω _x from the angular velocity sensor 20 in step (158). That is, the measured value of the roll angular velocity is acquired. From the data of 100 or more rollover tests, the angular velocity ω _x accompanying the rollover, that is, the roll angular velocity is generally ± 300 degrees / second.

The resulting lateral acceleration component A _y (t) is generally ± 20 g

It was found to be in the range. Lateral acceleration components A _y (t) and angular velocities ω _x exceeding the respective limit values are trimmed with the values in steps (154) and (160), respectively. For example, for an example where the related range is ± 20 g, the measured value of the lateral acceleration component A _y (t) less than −20 g is set to −20 g in step (154). The polarities of the lateral accelerometer 18 and the angular velocity sensor 20 are set such that the corresponding polarities of the angular velocity ω _x signal and the lateral acceleration component A _y signal are the same during roll generation. Generally the level at which the signal from the lateral accelerometer 18 is trimmed

Is set to the minimum value of either 20 g or the measurement range of the lateral accelerometer 18. Similarly, the level at which the signal from the angular velocity sensor 20 is trimmed

Is set to the minimum value of either 300 degrees / second or the measurement range of the angular velocity sensor 20.

およびフィルタリングされた

となる。フィルタリングされた測定値を使用することは、ロール判別アルゴリズムの誤開始を避けること、および測度アルゴリズム３００．１およびエネルギアルゴリズム３００．２による関連する判別処理を改善することにおいて有利である。フィルタ２２、２３は、例えば１０から１５ミリ秒の間の移動平均ウインドＴ_Ａｖｇを有する移動平均フィルタとして、迅速な信号応答とノイズ低減とを適切に妥協させる。例えば、プロセッサ２６は、以下に仮定するように、角速度ω_ｘ信号および横加速度成分Ａ_ｙ信号を、サンプリング速度２５００Ｈｚ（サンプリング周期ｄｔ＝０．４ミリ秒に対応）、ウインド１２．８ミリ秒で均一にサンプリングし、各信号についての移動平均が、収集された最新の３２のサンプルから計算される。移動平均の個々のサンプルは、通常均一に重みづけされるが、別の選択として不均一に重み付けすることもできる。 The raw data of the lateral acceleration component A _y and the angular velocity ω _x from the lateral accelerometer 18 and the angular velocity sensor 20 are respectively filtered by the respective filters 22 and 24 in steps (156) and (162), respectively. Was

And filtered

It becomes. Using filtered measurements is advantageous in avoiding false start of the role discrimination algorithm and improving the associated discrimination process by the measure algorithm 300.1 and energy algorithm 300.2. Filters 22, 23, for example, as a moving average filter having a moving average window T _Avg between 10 and 15 milliseconds, properly compromises rapid signal response and noise reduction. For example, as will be assumed below, the processor 26 generates an angular velocity ω _x signal and a lateral acceleration component A _y signal at a sampling rate of 2500 Hz (corresponding to a sampling period dt = 0.4 ms) at a window of 12.8 ms. Sampling uniformly, a moving average for each signal is calculated from the last 32 samples collected. The individual samples of the moving average are usually weighted uniformly, but can be weighted non-uniformly as another option.

およびフィルタリングされた

を出力する上記の移動平均フィルタよりも、大幅に低い有効カットオフ周波数、あるいは言い換えると大幅に大きい有効フィルタ時間定数を有する、関連するフィルタを用いてセンサ計測値をフィルタリングすることによって評価される。例えば、

は、ステップ（１６８）および（１７０）におけるそれぞれの移動平均フィルタによって、関連する角速度ω_ｘおよび横加速度成分Ａ_ｙそれぞれの計測生データをフィルタリングして得られ、この各移動平均フィルタは関連するウインド幅

例えば約４秒を有する。 In general, the lateral accelerometer 18 and the angular velocity sensor 20 may exhibit offset and / or drift errors (commonly referred to as sensor offset errors), and associated roll detection errors may occur unless this is compensated for. . This sensor offset error is filtered

And filtered

Is evaluated by filtering the sensor measurements with an associated filter that has a significantly lower effective cut-off frequency, or in other words, a significantly larger effective filter time constant, than the moving average filter described above. For example,

Is obtained by filtering the measured raw data of the respective angular velocity ω _x and lateral acceleration component A _y by the respective moving average filters in steps (168) and (170), each moving average filter being associated with the associated window. width

For example, having about 4 seconds.

の値は測度アルゴリズム３００．１およびエネルギアルゴリズム３００．２の両方が開始されていないときにのみ更新され、これは関連する両フラグONGOING_EVENT_FLAGs（進行イベントフラグ）、すなわちONGOING_MEASURES_EVENT_FLAG（進行メジャーズイベントフラグ）およびONGOING_ENERGY_EVENT_FLAG（進行エネルギイベントフラグ）の両方が設定されていないことで示される。したがって、ステップ（１６６）において、比較的長時間のフィルタリングを施された

の値は、関連する横加速度成分Ａ_ｙおよび角速度ω_ｘが、関連するセンサオフセット値と大幅に異なる可能性のある時間間隔においては、更新されない。 Filtered from step (164)

The value of is updated only when both the measure algorithm 300.1 and the energy algorithm 300.2 are not started, which are both related flags ONGOING_EVENT_FLAGs (progression event flags), namely ONGOING_MEASURES_EVENT_FLAG (progression majors event flag) and ONGOING_ENERGY_EVENT_FLAG It is indicated that both (progress energy event flag) are not set. Therefore, in step (166), a relatively long filtering was performed.

Is not updated in time intervals where the associated lateral acceleration component A _y and angular velocity ω _x can be significantly different from the associated sensor offset value.

をそれぞれ減じることにより補償した、フィルタリングされた

およびフィルタリングされた

を使用することによって、対応する補償横加速度成分

および補償角速度

がそれぞれ求められる。 In FIG. 4, the collection and processing of the lateral acceleration component A _x is shown before the collection and processing of the angular velocity ω _x , but the relative order may be reversed, and these operations are performed in parallel. It should be understood that it may be performed.
In each of the measure algorithm 300.1, energy algorithm 300.2, and safety algorithm 200, the corresponding sensor offset, i.e.

Filtered, compensated by subtracting each

And filtered

By using the corresponding compensated lateral acceleration component

And compensation angular velocity

Is required.

の絶対値が第３の加速度閾値

よりも大きい場合に、次のステップ（２０８）において、ACCELERATION_SAFING_EVENT_FLAGが設定される。そうでない場合には、ステップ（２０４）から、ステップ（２０２）に戻り処理が反復される。 Referring to FIG. 5, the safety algorithm 200 begins at step (202) where both relevant flags SAFING_EVENT_FLAGs (safety event flags), namely ACCELERATION_SAFING_EVENT_FLAG (acceleration safety event flag) and ROLL_SAFING_EVENT_FLAG (roll safety event flag). ) Is reset first. If either measure algorithm 300.1 or energy algorithm 300.2 is then started, as can be seen in step (204), either the ONGOING_EVENT_FLAGs flag (ie, ONGOING_MEASURES_EVENT_FLAG or ONGOING_ENERGY_EVENT_FLAG) is set In the next step (206),

Is the third acceleration threshold

If greater than, in the next step (208), ACCELERATION_SAFING_EVENT_FLAG is set. Otherwise, the process is repeated from step (204) to step (202).

の絶対値が第３の角速度閾値

よりも大きい場合に、ステップ（２１２）でROLL_SAFING_EVENT_FLAGが設定される。これに次いで、あるいはステップ（２１０）から、ステップ（２０４）に戻り反復処理される。したがって、測度アルゴリズム３００．１およびエネルギアルゴリズム３００．２の少なくとも１つに入った後で、かつ両者から出る以前に、安全化アルゴリズム２００に関連する横加速度および角速度の条件が満たされると、それが必ずしも同時でなくても、それぞれの関連する安全化イベントフラグSAFING_EVENT_FLAGsが次に設定されて、測度アルゴリズム３００．１またはエネルギアルゴリズム３００．２のいずれかによるロール状態の検出に応じて１つまたは複数の安全拘束アクチュエータの展開が可能となる。それぞれのSAFING_EVENT_FLAGsは個別に設定、またはラッチされるが、両者は同時にリセットされると共に、測度アルゴリズム３００．１またはエネルギアルゴリズム３００．２に応じて１つまたは複数の関連する安全拘束アクチュエータ３０を作動させるためには、両方が設定されていなくてはならない。 Following step (208) or following step (206), at step (210),

Is the third angular velocity threshold

Is greater than ROLL_SAFING_EVENT_FLAG is set in step (212). Following this, or from step (210), it returns to step (204) and is iteratively processed. Therefore, after entering at least one of the measure algorithm 300.1 and the energy algorithm 300.2 and before exiting from both, if the lateral acceleration and angular velocity conditions associated with the safety algorithm 200 are met, Although not necessarily simultaneous, each associated safety event flag SAFING_EVENT_FLAGs is then set to one or more depending on the detection of the roll condition by either the measure algorithm 300.1 or the energy algorithm 300.2. The deployment of the safety restraint actuator becomes possible. Each SAFING_EVENT_FLAGs is set or latched individually, but both are reset at the same time and activate one or more associated safety restraint actuators 30 depending on the measure algorithm 300.1 or energy algorithm 300.2 For this to happen, both must be set.

よりも大きい

の絶対値と、開始時刻に続く第２ポイント時刻における第３の角速度閾値

よりも大きい

の絶対値の、少なくとも一方に応答するようにしてもよく、この場合に開始時刻とは、関連する測度アルゴリズム３００．１またはエネルギアルゴリズム３００．２について、関連する入口基準が満たされる時刻であり、開始時刻に続く、第１ポイントおよび第２ポイント時刻は互いに任意である。例えば、エネルギアルゴリズム３００．２は、エネルギアルゴリズム３００．２の開始時刻に続くある時刻における、第３の加速度閾値

よりも大きな

だけに応答して「安全化」させることもできる。 As an alternative, by adapting the safety algorithm 200 to incorporate only one of the aforementioned flags and only the relevant criteria, the safety criteria can be either the measure algorithm 300.1 or the energy algorithm 300.2. The third acceleration threshold value at the first point time following the start time

Bigger than

And the third angular velocity threshold at the second point time following the start time

Bigger than

In which case the start time is the time at which the relevant entry criteria are met for the associated measure algorithm 300.1 or energy algorithm 300.2; The first point time and the second point time following the start time are arbitrary. For example, the energy algorithm 300.2 may include a third acceleration threshold at a time following the start time of the energy algorithm 300.2.

Bigger than

It can also be made "safe" in response to only.

（例えば１２秒）を設けてリセットしてもよい。 The rollover detection system 10 may be adapted to improve reliability by implementing the safety algorithm 200 using a microprocessor separate from that for implementing the measure algorithm 300.1 or energy algorithm 300.2. Well, in this case, if the safety algorithm 200 does not recognize the progress event flags ONGOING_EVENT_FLAGs, instead of resetting the safety event flag SAFING_EVENT_FLAGs according to these flags, one or more safety actuators 30 are deployed. From the time when one of the safety criteria was last met, e.g., so that the safety condition remains operational until both algorithms are finished or until both algorithms have ended.

(For example, 12 seconds) may be provided and reset.

  Each of the measure algorithm 300.1 and the energy algorithm 300.2 operates according to the overall algorithm structure shown in FIG. 6, and each of these algorithms is designated by the reference numeral 300. We will refer to individual algorithms by attaching a decimal number to a particular reference number. For example, the general overall process is represented by reference number 300, the measure algorithm is 300.1, and the energy algorithm is 300.2. As another example, the general algorithm calculation step is represented by reference number 326, while reference number 326.1 specifically represents the algorithm calculation step of measure algorithm 300.1, and reference number 362.2 represents energy algorithm 300.2. Represents the algorithm calculation step. The specific equations associated with a particular algorithm step are shown in tabular form in FIGS. 8a-8c for each algorithm, and the associated parameters and example values are shown in tabular form in FIGS. 9a, 9b.

  Referring to FIG. 6, the entire roll processing algorithm starts from step (302), and the corresponding progress event flag ONOING_EVENT_FLAG is reset. Setting this ONGOING_EVENT_FLAG indicates that the entry criterion of the roll processing algorithm is satisfied, the corresponding exit criterion is not satisfied, and therefore the associated algorithm is active. Then, in step (150), according to the data collection and processing algorithm 150 described above, the relevant data used in this algorithm is collected and processed. Then, in step (304), if ONGOING_EVENT_FLAG is not set, that is, if data from a potential roll event is not being processed, indicating that no roll event has occurred in vehicle 12, step (306) In step 308, a set of entrance criteria is evaluated and compared to an associated threshold, and if the entrance criteria are met, ONGOING_EVENT_FLAG is set in step (308), and in step (310), for example, various algorithm-related actions. The algorithm is started by initializing the static variables.

Alternatively, from step (304), if the progress event flag ONGOING_EVENT_FLAG is set to indicate that data from a potential roll event is being processed, then in step (312) the associated time measure, eg The sample count is updated and, in step (400), the newly collected data is evaluated to determine if the sensor (ie, lateral accelerometer 18 or angular velocity sensor 20) needs to be recalibrated. The process associated with step (400) is shown in FIG. 7 and will be described in more detail below.
From step (400), if one or more sensors require recalibration, in step (314), the one or more sensors that require recalibration are recalibrated. For example, both the lateral accelerometer 18 and the angular velocity sensor 20 can be tested and can be calibrated such that a known stimulus is applied to the sensor and the corresponding sensor output matches the known stimulus. For example, if the lateral accelerometer 18 includes a mass element suspended by a beam of a spring element by micromachining, an amount corresponding to a reference acceleration level by applying an electrostatic field between the mass element and the housing. Only bend the beam.

  A calibration factor is then calculated so that the output from the strain sensing element matches the reference acceleration level. In step (316), if the process of step (314) indicates that one or more sensors are faulty, for example, whether or not a test stimulus is applied to the sensors, the output is substantially changed. If not, in step (318) a fault condition is set and an alarm device, for example a light, is activated to alert the driver of the vehicle 12 and disable the deployment of the safety restraint actuator 30 by the rollover detection system. . Alternatively, from step (316), i.e., if both lateral accelerometer 18 and angular velocity sensor 20 have not failed, in step (320) both flags ONGOING_EVENT_FLAGs, i.e. ONGOING_MEASURES_EVENT_FLAG and ONGOING_ENERGY_EVENT_FLAG, It is reset in response to the calibration being performed, and a new process is repeated from step (150).

の結果ステップ（３２２）で退出し、ステップ（３２２）で退出した直後に、例えばアルゴリズムの次回反復中にステップ（３０６）で再び入る場合には、ステップ（３２２）においてｐ回（例えばｐ＝３）連続で退出した後に、上述のようにステップ（３２４）からの処理が継続されて、センサが診断され、必要な場合には再較正される。あるいはステップ（３２４）から、関連する進行フラグONGOING_EVENT_FLAG、すなわちONGOING_MEASURES_EVENT_FLAGまたはONGOING_ENERGY_EVENT_FLAGがステップ（３２０）でリセットされて、ステップ（１５０）から新しい処理が反復される。 Alternatively, from step (400), if no sensors need to be recalibrated, the exit criteria are evaluated in step (322) until the time when the entrance criteria in step (306) are re-filled and the algorithm is re-entered. It is determined whether to exit the algorithm. From step (322), if the exit criteria are met, in step (324) the algorithm is energy algorithm 300.2, and energy algorithm 300.2 is continuously registered in step (306), and Then timeout

As a result of exiting at step (322) and immediately after exiting at step (322), for example, when entering again at step (306) during the next iteration of the algorithm, at step (322) p times (eg, p = 3) ) After exiting continuously, processing from step (324) continues as described above, and the sensor is diagnosed and recalibrated if necessary. Alternatively, from step (324), the associated progress flag ONGOING_EVENT_FLAG, ie, ONGOING_MEASURES_EVENT_FLAG or ONGOING_ENERGY_EVENT_FLAG is reset at step (320) and the new process is repeated from step (150).

  Alternatively, from step (322), the time measure from either step (310) or step (312) is specified if the algorithm was registered at step (306) and not exited at step (322). Depending on the value of, the associated algorithm calculation is performed for a particular iteration of the algorithm. Then, in step (330), if the relevant algorithm detection criteria are met in a particular iteration of the algorithm and the safety event flag SAFING_EVENT_FLAG (s), ie ACCELERATION_SAFING_EVENT_FLAG and ROLL_SAFING_EVENT_FLAG, is set in step (340) In step (350), a roll event is detected and the associated safety restraint actuator 30 is activated. Conversely, if the algorithm detection criteria are not met in step (330), or in step (340), all the safety event flags are unset and the associated safety criteria are either the measure algorithm 300.1 or the energy algorithm. If not satisfied at some time during the execution of 300.2, the next iteration starting at step (150) is continued.

Both the measure algorithm 300.1 and the energy algorithm 300.2 depend on measurements of the lateral acceleration component A _y and the longitudinal angular velocity ω _x from the data collection and processing algorithm 150, but other variables associated with each algorithm. And parameters are independent of each other, as well as relevant entry criteria in step (306), algorithm initialization in step (310), relevant exit criteria in step (322), algorithm computation in step (326), and step ( The algorithm criteria in 330) are also independent, examples of all of which are shown in detail in FIGS. 8a, 8b, 8c, 9a and 9b. For example, each algorithm determines a time measure from the start and calculates a roll angle measure by integrating the measured values of the longitudinal angular velocity ω _x , as well as their respective roll angle measures. Time measures are independent of each other. Both measure algorithm 300.1 and energy algorithm 300.2 assume that the vehicle is initially horizontal (ie, θ (t _entrance ) = 0) when processing by the respective algorithm begins.

の絶対値が、関連する時間間隔

の間、第４の角速度閾値

を連続的に超える場合には、ステップ（４１０）において角速度センサ２０の再較正の信号が出される。あるいは、ステップ（４１２）、（４１４）、（４１６）、（４１８）および（４２０）において、測度アルゴリズムからのロール角の絶対値θ^Ｍか、またはエネルギアルゴリズム３００．２からのロール角θ^Ｅのいずれかが、関連する時間間隔

の間、ロール角閾値

を超える場合には、ステップ（４１０）において角速度センサ２０の再較正の信号が出される。あるいはステップ（４２２）において、角速度センサ２０の再較正の信号は出されない。ステップ（４２４）、（４２６）、（４２８）および（４３０）において、フィルタリングされた

の絶対値が、第４の横加速度閾値

を、関連する時間間隔

の間、連続的に超える場合には、ステップ（４３２）において横加速度計１８の再較正の信号が出される。そうでない場合には、ステップ（４３４）において、横加速度計１８の再較正の信号は出されない。ステップ（４１０）または（４３２）のいずれかで再較正の信号が出された場合には、上記のようにステップ（３１４）から処理が継続される。そうでない場合には、センサ再較正の信号は出されず、上記のようにステップ（３２２）から処理が継続される。 A process 400 for determining whether either the lateral accelerometer 18 or the angular velocity sensor 20 requires recalibration is shown in FIG. Filtered in steps (402), (404), (406) and (408)

Absolute value of the associated time interval

The fourth angular velocity threshold during

Is continuously exceeded, a signal for recalibration of the angular velocity sensor 20 is issued in step (410). Alternatively, in steps (412), (414), (416), (418) and (420), the absolute value of the roll angle θ ^M from the measure algorithm or the roll angle θ ^E from the energy algorithm 300.2 Either is the relevant time interval

Roll angle threshold during

Is exceeded, a signal for recalibration of the angular velocity sensor 20 is issued in step (410). Alternatively, in step (422), a signal for recalibration of the angular velocity sensor 20 is not issued. Filtered in steps (424), (426), (428) and (430)

Is the fourth lateral acceleration threshold value.

The relevant time interval

If it exceeds continuously during step 432, a signal for recalibration of the lateral accelerometer 18 is issued in step (432). Otherwise, in step (434), no signal for recalibration of the lateral accelerometer 18 is issued. If a recalibration signal is issued in either step (410) or (432), processing continues from step (314) as described above. Otherwise, no sensor recalibration signal is issued and processing continues from step (322) as described above.

  The measure algorithm 300.1 will now be discussed in more detail with reference to FIGS. 6, 8a-8c, and FIGS. 9a, 9b, but the step numbers in FIG. A subscript such as “1” is attached. The progress event flag ONGOING_EVENT_FLAG of the measure algorithm 300.1 is called the progress measure event flag ONGOING_MEASURES_EVENT_FLAG. When the entrance criterion is satisfied in step (306.1), it is set in step (308.1), and step (322.1). If the exit criterion is satisfied in step (320.1), it is reset. The progress measure event flag ONGOING_MEASURES_EVENT_FLAG may correspond to a particular location in the memory 28 of the associated processor 26, for example, for implementing the measure algorithm 300.1. Following the step (322.1), upon entering the algorithm, after that, the exit criterion of the measure algorithm 300.1 is met in step (322.1) or a roll event is detected and the safety restraint actuator 30 is deployed. Until then, measure algorithm 300.1 will not leave. Further, after the measure event exit criteria are met and exit from the measure algorithm 300.1, the measure algorithm 300.1 may reenter if the associated measure event entry criteria are subsequently met.

の絶対値が第１の加速度閾値

より大きい、すなわち

であることである。ある特定の車両の例としては、実際のロールオーバのデータに基づいて、第１の加速度閾値は、約１．４ｇに設定された。この閾値は、測度アルゴリズム３００．１の他のパラメータ値と同様に、一般に特定の関連する車両の特性または車両のクラスに依存すること、および特定のロールオーバ検出システム１０に用いる特定の値は、関連する車両の１２の特性または車両のクラスに応じて調節することによって判別能力を改善することができることに留意すべきである。 The entry criterion of the measure algorithm 300.1 in step (306.1) is, for example:

Is the first acceleration threshold

Greater than

It is to be. As an example of a particular vehicle, based on actual rollover data, the first acceleration threshold was set to about 1.4 g. This threshold, like other parameter values of the measure algorithm 300.1, generally depends on a particular relevant vehicle characteristic or vehicle class, and the particular value used for a particular rollover detection system 10 is: It should be noted that the discriminating ability can be improved by adjusting according to the 12 characteristics of the vehicle concerned or the class of vehicle.

In step (310.1), when the measure algorithm 300.1 is first entered following step (308.1), the measure algorithm 300.1 is initialized. The number of event samples n ^M, the value θ ^M (n ^M −1) of the angular position, and the measure function R (n ^M −1) are initialized to, for example, zero values. The sample time t ^M (−1) immediately before the event entry time is initialized to the value of the measure event entry time t ^M (0), which is initialized at the current time t, and the time interval Δt from the algorithm entry. ^M (0) is initialized to a zero value. The superscript “M” used here represents a variable related to the measure algorithm 300.1.

が、現行時刻ｔに等しく設定され、測度イベント時間Δｔ^Ｍが測度イベント入口時刻ｔ^Ｍ（０）から現行時刻

までの時間間隔として、以下のように計算される。

ステップ（３２２．１）における、測度アルゴリズム３００．１の出口基準は、例えばアルゴリズム入口からの時間間隔

が、時間間隔閾値

より大きいこと、すなわち

となることである。特定型式の車両の例については、実際のロールデータに基づいて、時間間隔閾値

は約１６５ミリ秒に設定された。測度アルゴリズム３００．１から出ると、進行測度イベントフラグONGOING_MEASURES_EVENT_FLAGはステップ（３２０．１）でリセットされ、これによって、その後にステップ（３０６．１）において入口基準が満たされるまでの間、ステップ（３１０．１）で測度アルゴリズム３００．１に関連する変数が初期化される。 During subsequent iterations of the measure algorithm 300.1, the progress measure event flag is set in step (304.1), then the number of event samples n ^M is incremented in step (312.1) and the associated current sample time.

Is set equal to the current time t, and the measure event time Δt ^M is the current time from the measure event entry time t ^M (0).

Is calculated as follows.

The exit criterion of the measure algorithm 300.1 in step (322.1) is, for example, the time interval from the algorithm entry

Is the time interval threshold

Greater than, i.e.

It is to become. For examples of specific types of vehicles, the time interval threshold is based on actual roll data.

Was set to about 165 milliseconds. Upon exiting from the measure algorithm 300.1, the progress measure event flag ONGOING_MEASURES_EVENT_FLAG is reset in step (320.1), so that step (310. In 1), variables associated with the measure algorithm 300.1 are initialized.

ここで、積分時間ステップｄｔは、現行反復における時刻

と、前回反復における時刻との差

として求め、この差はサンプリング速度が一様であれば、以下のように一定値となる。

また、

は以下の式で求められる。

In step (322.1), if the exit criteria are not met, the algorithm calculation is updated in step (326.1) for a particular iteration of measure algorithm 300.1 as follows.
First, by integrating the compensated signed angular velocity value, the angular position θ ^M is estimated as follows:

Where the integration time step dt is the time in the current iteration

And the time difference from the previous iteration

If the sampling rate is uniform, this difference becomes a constant value as follows.

Also,

Is obtained by the following equation.

測度関数の第１項は、ダンピング項であり、前回値Ｒ(n^M-1)にダンピング係数

を乗じた積からなる。ダンピングのレベルは、特定の車両型式に依存する定数τによって決まる。例をあげると、特定型式の車両についてのロールオーバ試験データに基づき、τの値は約４００秒とした。ダンピング項は、

または

の値が継続的に大きくない事象に対しては、結果として得られる良度指数ＦＯＭを確実に低下させる働きをする。 A measure function R is then estimated, which is used to calculate a figure-of-merit FOM. The measure function R is obtained as follows.

The first term of the measure function is the damping term, and the damping factor is added to the previous value R (n ^M -1).

It consists of the product of. The level of damping is determined by a constant τ that depends on the specific vehicle type. As an example, the value of τ was about 400 seconds based on rollover test data for a specific type of vehicle. The damping term is

Or

For events where the value of is not continuously large, it serves to reliably reduce the resulting goodness index FOM.

の現行サンプルとして以下のように求められる。

一般に、力と加速度はニュートンの第２法則の関係がある（Ｆ＝Ｍ・Ａ）。力測度Ｆ^*は、力と加速度のベクトル特性の説明を必要とするような、厳密な力測度とする必要は必ずしもなく、少なくとも車両１２に作用する反力Ｆと相関のある測度とする。通常の車両のロール事象において、

は、タイヤまたは車輪リムに作用する横方向力によって生じる。この横方向力は、車両重量中心の回りの回転トルクを発生させて、最終的にロールオーバを引き起こす力と同じものである。

は、必ずしも実際の反力Ｆに対する完全な測度とはならない。 The remaining term added to the first term of the measure function R is the product of three measures: force measure F ^* , rotational kinetic energy measure KE ^* , and potential energy measure PE ^* . The force measure F ^* is

As a current sample, the following is required.

In general, force and acceleration are related to Newton's second law (F = MA). The force measure F ^* is not necessarily a strict force measure that needs to explain the vector characteristics of force and acceleration, and is a measure that is correlated with at least the reaction force F acting on the vehicle 12. In normal vehicle roll events,

Is caused by a lateral force acting on the tire or wheel rim. This lateral force is the same as the force that ultimately generates rollover by generating a rotational torque about the vehicle weight center.

Is not necessarily a perfect measure for the actual reaction force F.

では、例えばタイヤまたはその他のダンピング要素内の減衰力によるか、またはサスペンションシステムの運動による、非剛体運動の影響は、必ずしも説明できない。しかしながら

は経験則として、非剛体運動を除いては、小さな角度に対しては、車両１２をロールさせる反力Ｆに比例する。高速ロールオーバ試験またはトリップ式ロールオーバ試験によると、

は、角速度センサ２０によって有効な補償角速度が観察される約２０ミリ秒前に、有効となる。力測度Ｆ^*は、

に対して線形関係で示してあるが、力測度Ｆ^*は、

に対して、その他の（線形以外の）関数または（１以外の）累乗とすることもできることを理解すべきである。 For example

Then, the influence of non-rigid motion, for example, due to damping forces in the tire or other damping elements or due to the motion of the suspension system, cannot always be explained. However

As a rule of thumb, except for non-rigid motion, it is proportional to the reaction force F that rolls the vehicle 12 for small angles. According to the high speed rollover test or trip type rollover test,

Becomes effective approximately 20 milliseconds before the effective compensation angular velocity is observed by the angular velocity sensor 20. The force measure F ^* is

The force measure F ^* is

However, it should be understood that other (non-linear) functions or powers (other than 1) can be used.

で求められる。一般に回転運動エネルギ測度ＫＥ^*は、車両の回転運動エネルギと相関がある。例えば、

のように、回転運動エネルギ測度ＫＥ^*は、比例定数をＩ_ｘ／２として車両１２の回転運動エネルギと比例する。しかし、回転運動エネルギ測度ＫＥ^*を、別の形式で表すことも可能である。例えば、

の２乗以外の累乗を用いて、

から回転運動エネルギ測度ＫＥ^*を構成することも可能であり、あるいは回転運動エネルギ測度ＫＥ^*を

の別の関数とすることも可能である。 The rotational kinetic energy measure PE ^* is

Is required. In general, the rotational kinetic energy measure KE ^* is correlated with the rotational kinetic energy of the vehicle. For example,

As described above, the rotational kinetic energy measure KE ^* is proportional to the rotational kinetic energy of the vehicle 12 with the proportionality constant being I _x / 2. However, the rotational kinetic energy measure KE ^* can also be expressed in other forms. For example,

Using a power other than the square of

It is also possible to configure a rotational kinetic energy measure KE ^* from or rotational kinetic energy measure KE ^*

It is also possible to use another function.

を単独で用いる場合よりも、より迅速にロールオーバを予測する測度となる。これはまた

の最終値の予測測度ともなるが、それは

から推定された大きな横方向力は、通常約２０ミリ秒後に

の増加として表れることが観察されているからである。さらに、例えば補償角速度の２乗を用いることにより、補償角速度の重み付けを、補償横加速度成分よりも大きくすると、結果的に得られる良度指数ＦＯＭに対する、実際の補償角速度の影響を強めることになる。 The product of force measure F ^* and rotational kinetic energy measure KE ^* is

It is a measure that predicts rollover more quickly than when using a single. This is also

Which is also a predictive measure of the final value of

The large lateral force estimated from is usually about 20 milliseconds later

This is because it has been observed that it appears as an increase in. Further, for example, by using the square of the compensation angular velocity, if the weighting of the compensation angular velocity is made larger than the compensation lateral acceleration component, the effect of the actual compensation angular velocity on the resulting merit index FOM is strengthened. .

の和として次のように求められる。

定数θ_０は、特定の車両によって異なる。例えば、ある型式の車両についてのロールオーバ試験のデータに基づけば、θ_０の値は約０．１度である。この定数項は、

および

の両方の信号が、あるロール事象に対して同じ極性を有するように極性が与えられていると仮定して、これらの信号のいずれかと同じ同じ符号にされている。位置エネルギ測度ＰＥ^*を測度関数の積項に含めると、結果的に得られる良度指数ＦＯＭに与えるロール運動の影響を強め、例えば関連するアクチュエータ起動時間（ＴＴＦ）が通常１４０から２３０ミリ秒である、中速度のロール事象に対して、その絶対値を増大させる。（この範囲の上下限は、車両の特性に応じて２０％程度拡張することもあり、また異なる型式の車両に対してさらに異なる可能性もある。）力測度Ｆ^*および回転運動エネルギ測度ＫＥ^*と比較して、位置エネルギ測度ＰＥ^*は、比較的重要度が低く、小規模なロールオーバ検出システム１０においては無視する（例えばＰＥ^*＝１に設定する）ことも可能である。しかし位置エネルギ測度ＰＥ^*は、中程度のアクチュエータ起動時間を示すロール事象の場合の要素として利点があると思われる。 The potential energy measure PE ^* is the current sample of constant and angular position

Is calculated as follows.

The constant θ ₀ varies depending on the specific vehicle. For example, based on rollover test data for a certain type of vehicle, the value of θ ₀ is about 0.1 degrees. This constant term is

and

Assuming that both signals are given the same polarity for a roll event, they have the same sign as either of these signals. Inclusion of the potential energy measure PE ^* in the product term of the measure function enhances the effect of the roll motion on the resulting figure of merit FOM, eg, the associated actuator start-up time (TTF) is typically 140 to 230 milliseconds. For a medium speed roll event, increase its absolute value. (The upper and lower limits of this range may be expanded by about 20% depending on the characteristics of the vehicle and may be further different for different types of vehicles.) Force measure F ^* and rotational kinetic energy measure KE ^* In comparison, the potential energy measure PE ^* is relatively insignificant and can be ignored (eg, set to PE ^* = 1) in the small rollover detection system 10. However, the potential energy measure PE ^* appears to be advantageous as a factor in the case of a roll event that exhibits moderate actuator activation time.

良度指数ＦＯＭは、関連するR(n^M)およびR(n^M-1)項の絶対値から計算され、この結果良度指数ＦＯＭは回転の方向には影響されない。

項は、時間についての測度関数Ｒの微分または勾配の測度となり、実際の勾配はこの項をサンプリング周期ｄｔ（一様にサンプリングされるシステムでは定数）で除して求められることになる。この勾配係数は、以下に述べる閾値関数と組み合わされて、ロールオーバ事象が検出されて、その結果、関連する1つまたは複数の安全拘束システム３０展開させるためには、良度指数ＦＯＭが時間と共に増加することを必要とさせる効果がある。 The goodness index FOM is obtained by the following equation.

The figure of merit index FOM is calculated from the absolute values of the associated R (n ^M ) and R (n ^M −1) terms, so that the figure of merit index FOM is not affected by the direction of rotation.

The term is a derivative of the measure function R with respect to time or a measure of the slope, and the actual slope is determined by dividing this term by the sampling period dt (a constant in a uniformly sampled system). This slope factor is combined with the threshold function described below so that a rollover event is detected and, as a result, the figure of merit FOM is timed to deploy the associated safety restraint system 30 or systems 30. There is an effect that needs to be increased.

の値について、良度指数ＦＯＭは次式で求めてもよい。

ステップ（３２２．１）のアルゴリズム計算に続いて、ステップ（３３０．１）で評価されたアルゴリズム検出基準は、例えば図８ｃに示すような複数の検出条件を含む。これらの条件のすべてが満たされる場合には、通常は測度イベント閾値を超えて、ロールオーバが発生すると考えられ、またステップ（３４０）において安全化アルゴリズムから関連する安全化基準が満たされる場合には、ステップ（３５０）において、関連する乗員の傷害を軽減するために、関連する１つまたは複数の安全拘束アクチュエータ３０が展開される。この検出基準は、特定の検出戦略にしたがって決められる。理想的には、検出基準は、関連する１つまたは複数の安全拘束システムの適切な展開によって乗員の傷害が軽減される程度の強度の、車両の内装と頭部の接触（つまり「頭部接触」）が発生するロール事象を検出することができて、その他の事象を無視できることである。 As an alternative, especially relatively small

The goodness index FOM may be obtained by the following equation.

Subsequent to the algorithm calculation in step (322.1), the algorithm detection criterion evaluated in step (330.1) includes a plurality of detection conditions as shown in FIG. 8c, for example. If all of these conditions are met, usually the measure event threshold is exceeded and a rollover is considered to occur, and if the relevant safety criteria are met from the safety algorithm in step (340). In step (350), the associated one or more safety restraint actuators 30 are deployed to reduce the associated occupant injury. This detection criterion is determined according to a specific detection strategy. Ideally, the detection criteria is a vehicle interior-head contact (ie, “head contact” that is strong enough to reduce occupant injury by appropriate deployment of the associated safety restraint system or systems. )) Can be detected and other events can be ignored.

  However, if such ideal performance cannot be achieved, the detection criteria can be adapted to be an appropriate compromise. For example, more than a related head contact time so that a significant roll event can be deployed quickly enough, that is, the associated safety actuator 30 can be deployed at a timing and speed that can reduce the risk of injury to the occupant. However, in order to detect within a sufficiently short period of time, in response to a critical rollover event that does not cause the vehicle to roll completely (eg curb trip or medium to high deceleration type roll event). It may be necessary to allow one or more safety restraint actuators 30 to deploy.

例えば特定の型式の車両に対して、イベントトリガーからの経過時間間隔が、特定の時間ウインド内に入るように、関連する最小および最大イベント時間を、

＝４０ミリ秒、

＝１６５ミリ秒とする。最小測度イベント時間

を制約することによって、非常に短時間の大きな横方向入力によって、偶然の検出が起こるのを防ぐと共に、観察された最短の頭部接触時間の条件を満たすように、十分に早く安全拘束システムを展開させることが可能になる。（頭部接触時間は、乗員の頭部が車両の内装と接触する時間である。）通常、重大な縁石走行または減速スレッド事象に対して、ロール判別アルゴリズムの入口時間は、ロール事象開始（すなわち、物理事象の開始）後、約２０ミリ秒である。そうすると、ロール判別アルゴリズムがエアバッグを展開し始めることができる最早時点は、ロール事象の開始後、約６０ミリ秒となる（入口時間＋約４０ミリ秒）。 As a first detection condition of step (330.1), it is tested as follows whether or not the measure event time Δt ^M is within the range of the measure event time (Δt ^M _min , Δt ^M _max ).

For example, for a particular type of vehicle, the associated minimum and maximum event times are set so that the elapsed time interval from the event trigger falls within a specific time window.

= 40 milliseconds,

= 165 milliseconds. Minimum measure event time

To prevent accidental detection from occurring in a very short, large lateral input and to ensure that the safety restraint system is fast enough to meet the conditions for the shortest head contact time observed. It becomes possible to develop. (Head contact time is the time when the occupant's head contacts the vehicle interior.) Typically, for critical curb travel or deceleration thread events, the roll discriminating algorithm entry time is the roll event start (ie, About 20 milliseconds after the start of the physical event). Then, the earliest point at which the roll discrimination algorithm can begin to deploy the airbag is approximately 60 milliseconds (entrance time + approximately 40 milliseconds) after the start of the roll event.

The fastest head contact time observed is on the order of 115 milliseconds after the occurrence of a roll event. Assuming that the associated data processing and safety restraint deployment (eg, airbag inflation) takes about 30 milliseconds, in these cases, the safety restraint actuator 30 is fully deployed about 90 milliseconds after the start of the roll event. Will be. The minimum activation time Δt ^min uses information provided by the signals from the lateral accelerometer 18 and the angular velocity sensor 20 as much as possible, while making it possible to make a deployment decision with sufficient time to avoid head contact in a serious event. To. The maximum start-up time Δt ^max reduces the vulnerability of the roll discrimination algorithm to chain events and resets the roll discrimination algorithm in the event of a rollover caused by the second event of two lateral events during the time interval. , Enabling the detection of the second “real” start event of the rollover. In step (330.1), if the measure event time Δt ^M is within the specified range, the first detection condition is satisfied and an additional detection criterion is evaluated in step (330.1). Otherwise, processing continues with the next iteration from step (150).

とが比較される。閾値関数

は、例えば以下の形式となる。

対応する第２の検出条件は以下のようになる。

例えば、特定の型式の車両についてのロールオーバ試験からのデータに基づいて、ＡおよびＢは、４０ミリ秒≦Δｔ^Ｍ＜９６ミリ秒に対して、Ａ＝６．４６×１０^１１（ｇ^２ｄｅｇ^６／ｍｓ・ｓ^４）、Ｂ＝−２．３４×１０^１３（ｇ^２ｄｅｇ^６／ｓ^４）、また９６ミリ秒≦Δｔ^Ｍ≦１６５ミリ秒に対して、Ａ＝２．５９×１０^１１（ｇ^２ｄｅｇ^６／ｍｓ・ｓ^４）、Ｂ＝−１．３６×１０^１３（ｇ^２ｄｅｇ^６／ｓ^４）とした。良度指数ＦＯＭおよび閾値関数

は、例えば両方とも単位が［ｇ^２ｄｅｇ^６／ｓ^４］である。 As a second detection condition in step (330.1), for the example vehicle platform, a goodness index FOM and sufficiently fast for substantially all events required by the above detection strategy. Threshold function that yields accurate discrimination time

Are compared. Threshold function

Is, for example, in the following format:

The corresponding second detection condition is as follows.

For example, based on data from a rollover test for a particular type of vehicle, A and B are A = 6.46 × 10 ¹¹ (g ² deg for 40 ms ≦ Δt ^M <96 ms. ⁶ / ms · s ⁴ ), B = −2.34 × 10 ¹³ (g ² deg ⁶ / s ⁴ ), and 96 ms ≦ Δt ^M ≦ 165 ms, A = 2.59 × 10 ¹¹ (G ² deg ⁶ / ms · s ⁴ ) and B = −1.36 × 10 ¹³ (g ² deg ⁶ / s ⁴ ). Goodness index FOM and threshold function

For example, both units are [g ² deg ⁶ / s ⁴ ].

について、それぞれ異なるパラメータ値、または異なる関数形を有する。例えば複数の線形セグメントを有する、多重セグメント閾値関数

は、ロール判別アルゴリズムの性能に対して有利であることがわかっている。上記の例示的閾値限界線は、０．８ミリ秒の時間ステップの間に１２５０Ｈｚの速度でサンプリングされたデータから設定された。閾値関数

は、良度指数ＦＯＭの計算における積分の残余項が異なるために、データサンプリング速度が変われば、変化する。一般に、閾値関数は、時間の関数、時間の分割関数、時間についてのテーブルルックアップのいずれかによって表すことができる。さらに通常、閾値関数は時間の関数、例えば開始から時間間隔の関数であるが、この閾値関数

は場合によっては一定、すなわち時間に対して一定である可能性もあることを理解すべきである。ステップ（３３０．１）において、良度指数ＦＯＭが閾値関数

を超える場合には、第２の検出条件が満たされ、ステップ（３３０．１）において、追加の検出基準が評価される。そうでない場合には、ステップ（１５０）から次の反復の処理が継続する。 In general, different types of vehicles have a threshold function

Each have a different parameter value or a different function form. Multi-segment threshold function, eg with multiple linear segments

Has been found to be advantageous for the performance of the role discrimination algorithm. The above exemplary threshold limit line was set from data sampled at a rate of 1250 Hz during a 0.8 millisecond time step. Threshold function

Changes when the data sampling rate changes because the integral residual term in the calculation of the figure of merit index FOM is different. In general, the threshold function can be represented by either a function of time, a time division function, or a table lookup for time. More usually, the threshold function is a function of time, for example, a function of time interval from start, but this threshold function

It should be understood that may be constant in some cases, i.e. constant over time. In step (330.1), the goodness index FOM is a threshold function.

If the second detection condition is exceeded, the second detection condition is satisfied and in step (330.1) additional detection criteria are evaluated. Otherwise, the process of the next iteration continues from step (150).

である。第３の検出条件は、例えばΔｔ＝４０ミリ秒で閾値関数ＦＯＭ^Thr（Δt）を超えるが、事象は減退しつつある（例えばＡ_ｙまたはω_ｘのいずれか、あるいは両方の絶対値が、減少している）場合の展開を防ぐためのものである。ステップ（３３０．１）において、良度指数ＦＯＭが時間に対して増大している場合には、第３の検出条件は満たされ、ステップ（３３０．１）において追加の検出基準が評価される。そうでない場合には、ステップ（１５０）から次の反復のための処理が継続される。 As a third detection condition of step (330.1), whether or not the absolute value of the goodness index FOM increases with time at the expected deployment of the safety restraint actuator 30 is as follows. test.

It is. The third detection condition exceeds the threshold function FOM ^Thr (Δt) at Δt = 40 milliseconds, for example, but the event is declining (eg, the absolute value of either A _y or ω _x or both decreases It is to prevent the deployment in the case. In step (330.1), if the goodness index FOM increases with time, the third detection condition is satisfied, and additional detection criteria are evaluated in step (330.1). Otherwise, the process for the next iteration continues from step (150).

の絶対値が、第２の加速度閾値

と以下のように比較される。

第４の検出条件は、誤った大きな補償角速度

信号を発生させるモードにおける、角速度センサ２０の故障によって、安全拘束アクチュエータ（群）３０の偶発的な展開が生じることを防止する。例えば、ドライブ路面上に横方向のタイヤスリップのないような正常な運転状態において、第２の加速度閾値

の値０．７ｇを超えることはないと考えられる。ステップ（３３０．１）において、

の絶対値が、第２の加速度閾値

よりも大きい場合には、第４の検出条件が満たされ、ステップ（３３０．１）において追加の検出基準が評価される。そうでない場合には、ステップ（１５０）から次の反復のための処理が継続される。 As a fourth detection condition of step (330.1), a compensated lateral acceleration component at the time of predicted deployment of the safety restraint actuator 30

Is the second acceleration threshold value.

And is compared as follows.

The fourth detection condition is an incorrect large compensation angular velocity

Accidental deployment of the safety restraint actuator (s) 30 due to failure of the angular velocity sensor 20 in the signal generating mode is prevented. For example, in a normal driving state where there is no lateral tire slip on the drive road surface, the second acceleration threshold value

It is considered that the value of 0.7 g is not exceeded. In step (330.1)

Is the second acceleration threshold value.

If greater than, the fourth detection condition is satisfied and an additional detection criterion is evaluated in step (330.1). Otherwise, the process for the next iteration continues from step (150).

の絶対値が、安全拘束アクチュエータ（群）３０の予想される展開時における、関連する第２のロール角速度閾値

と、以下のように比較される。

例えば、第２のロール角速度閾値

は、約５０度／秒である。第５の検出条件は、車両１２が安全拘束アクチュエータ（群）３０の展開時に、大きな角速度が発生していることを保証するものである。第２および第５の検出条件を組み合わせて、重大な側面衝突事象が、安全拘束アクチュエータ（群）３０を展開するのを防止する。第５の検出条件はまた、故障横加速度計１８が、誤った大きな横加速度信号を示して、安全拘束アクチュエータ３０を偶発的に展開するのを防止する。ステップ（３３０．１）において、

の絶対値が、第２の角速度閾値

よりも大きい場合には、第５の検出条件が満たされ、ステップ（３４０）から処理が継続する。そうでない場合には、ステップ（１５０）から次の反復のための処理が継続される。 As the fifth detection condition of step (330.1),

Is the associated second roll angular velocity threshold at the expected deployment of the safety restraint actuator (s) 30

Is compared as follows.

For example, the second roll angular velocity threshold

Is about 50 degrees / second. The fifth detection condition is to ensure that a large angular velocity is generated when the vehicle 12 is deployed with the safety restraint actuator (s) 30. The second and fifth detection conditions are combined to prevent a serious side impact event from deploying the safety restraint actuator (s) 30. The fifth detection condition also prevents the failed lateral accelerometer 18 from showing an erroneous large lateral acceleration signal and unintentionally deploying the safety restraint actuator 30. In step (330.1)

Is the second angular velocity threshold

If greater than, the fifth detection condition is satisfied, and the process continues from step (340). Otherwise, the process for the next iteration continues from step (150).

  The measure algorithm 300.1 described here has passed tests using a series of vehicle rollover test data and has proven to be able to reliably predict the final vehicle rollover. Roll events caused by large lateral accelerations can be predicted relatively quickly so that the measure algorithm 300.1 can be used for the type of rollover where head contact usually occurs most quickly. The airbag can be deployed before contact. In general, measure algorithm 300.1 includes relatively early rollover detection and associated safety restraint actuators for short and medium time roll events, similar to curb travel and high g lateral deceleration events. There is an advantage in realizing the relatively early start-up time (TTF) of (group) 30.

Thus, the rollover detection system 10 incorporating the measure algorithm 300.1 enables rollover airbag deployment time that matches the occupant head contact time while minimizing accidental deployment by the following steps: An improved identification method for vehicle rollover is realized.
It uses the measured lateral acceleration to predict the future (after 20-30 milliseconds) roll motion,
Combines lateral acceleration with angular velocity and full rotation angle to generate current rotation state and motion and rotation for roll events where the starting angle is less than about 20 degrees from horizontal without requiring initial vehicle angle information Generating a measure with a coercive function
Use vehicle-specific dynamic characteristics (extracted from rollover test data) in combination with initial vehicle response measurements to predict the eventual rollover before such events become apparent It is to be.

  Referring to FIG. 10, four different test conditions, named Test A, Test B, Test C, and Test D, are tabulated for purposes of explaining and comparing the measure algorithm 300.1 and the energy algorithm 300.2. (Energy algorithm 300.2 is described in further detail below). Test A and Test B are spiral tests, which show that the energy algorithm 300.2 shows a faster rollover detection than the measure algorithm 300.1, and tests C and D are the deceleration thread test For these, the measure algorithm 300.1 shows a faster rollover detection than the energy algorithm 300.2. In tests A and D, the vehicle rolled over, but in tests B and C, no rollover occurred and the maximum roll angles were 37 degrees and 34 degrees, respectively. The initial vehicle speed, average vehicle deceleration, associated detection and event times are also shown in FIG. 10, where the head contact time is the time when the head of the occupant (dummy) actually collided with the interior of the vehicle. is there.

を示してある。テストＤでは、車両がロールオーバを起こし、テストＣでは、最大回転角が約３４度まで達した。 Referring to FIGS. 11 a-11 d, filtered roll speed (angular velocity) from acceleration sensor 20, roll angle, and filtered lateral from lateral accelerometer 18 according to the conditions displayed in FIG. 10 for tests AD. Each acceleration is shown as a function of time.
Referring to FIG. 12, the calculated merit index for tests C and D is the measure event time Δt ^M , the measure algorithm in the actual thread deceleration test for the specific type of vehicle shown in the table of FIG. Plotted as a function of time from the start of 300.1. FIG. 12 shows the associated threshold function for this particular type of vehicle.

Is shown. In test D, the vehicle rolled over, and in test C, the maximum rotation angle reached approximately 34 degrees.

と、関連する閾値関数

を結合させることによって、車両のロールオーバが発生したテストＤにおけるロールオーバ事象の開始後、起動時間（ＴＴＦ）を９８ミリ秒とすることが可能となり、この時間は関連する頭部接触時間１９６ミリ秒よりも時間的に大幅に短く、これによって関連する１つまたは複数の安全拘束アクチュエータ３０を展開するための時間として９８ミリ秒がもたらされた。関連する安全化アルゴリズム２００の安全化基準は、ロール事象の開始後２６ミリ秒で満たされ、これは測度アルゴリズム３００．１が検出基準を満たすよりも時間的に大幅に早かった。これと比較して、以下に述べるエネルギアルゴリズム３００．２の検出基準は、テストＤの事象に対してはロール事象の開始後５９４ミリ秒までは満たされず、これは関連する頭部接触時間よりも大幅に遅れており、これによってテストＤのロール事象に対する測度アルゴリズム３００．１の利点が示された。 Merit index calculated by the measure algorithm 300.1 described here

And the associated threshold function

Can be used to provide a start-up time (TTF) of 98 milliseconds after the start of a rollover event in test D, where a vehicle rollover has occurred, which is associated with an associated head contact time of 196 milliseconds. Much shorter in time than seconds, this resulted in 98 milliseconds to deploy the associated one or more safety restraint actuators 30. The related safety algorithm 200 safety criteria was met 26 milliseconds after the start of the roll event, which was significantly faster in time than the measure algorithm 300.1 met the detection criteria. In comparison, the detection criteria of the energy algorithm 300.2 described below are not met for the test D event until 594 milliseconds after the start of the roll event, which is greater than the associated head contact time. Significantly late, which showed the benefit of the measure algorithm 300.1 over the roll event for test D.

  The energy algorithm 300.2 is described in more detail with reference to FIGS. 6, 8a-c, and 9a, 9b, where the step of FIG. It is related. The progress event flag ONGOING_EVENT_FLAG for the energy algorithm is called ONGOING_ENERGY_EVENT_FLAG and is set in step (308.2) when the entry criteria in step (306.2) are met, and the exit criteria is met in step (322.2). Is reset in step (320.2). For example, ONGOING_ENERGY_EVENT_FLAG may correspond to a particular location in the memory 28 of the associated processor 26 that implements the energy algorithm 300.2. Subsequent to step (306.2), after entering energy algorithm 300.2, in step (322.2) the energy event exit criteria is met or roll event 300.2 is detected and the safety restraint actuator Until 30 is deployed, it will not leave. Further, after the energy event exit criteria are met and exit from the energy algorithm 300.2, the energy algorithm 300.2 can be re-entered if the associated energy event entry criteria are subsequently met.

の絶対値が、第１の加速度閾値

よりも大きいか、または（ＯＲ）

の絶対値が第１のロール速度閾値

よりも大きいことである。すなわち、次式である。

特定の型式の車両を例にあげると、実際のロールオーバのデータに基づいて、第１の加速度閾値

を約１．４ｇ（測度アルゴリズム３００．１に関して）に設定し、第１の回転速度閾値

を約１９度／秒に設定した。この閾値は、エネルギアルゴリズム３００．２の他のパラメータ値と同様に、通常関連する特定の車両１２の特性または車両のクラスに依存し、かつ特定のロールオーバ検出システムに使用する特定の値は、関連する車両１２の性質あるいは車両のクラスに応じて判別を向上させるために調整することができることはいうまでもない。 The energy algorithm 300.2 uses the angular velocity ω _x signal from the angular velocity sensor 20 to identify the roll state of the vehicle, and is necessary to produce a complete roll with the overall energy (rotational motion and potential energy) of the vehicle 12. Compare the energy.
In step (306.2), the entrance criterion of the energy algorithm 300.2 is, for example,

Is the first acceleration threshold value

Is greater than or (OR)

Is the first roll speed threshold value

Is bigger than that. That is, the following equation.

Taking a specific type of vehicle as an example, the first acceleration threshold is based on actual rollover data.

Is set to about 1.4 g (with respect to the measure algorithm 300.1) and the first rotational speed threshold

Was set to about 19 degrees / second. This threshold, like other parameter values of the energy algorithm 300.2, depends on the characteristics of the particular vehicle 12 or class of vehicles normally associated, and the particular value used for a particular rollover detection system is: Of course, adjustments can be made to improve discrimination depending on the nature of the vehicle 12 or class of vehicle involved.

が初期化、つまりゼロに設定される。またイベント入口時刻の直前のサンプル時刻

は、現在時刻ｔの値に初期化されている、エネルギイベント入口時刻

に初期化され、かつアルゴリズムに入ってからの時間間隔

が、ゼロ値に初期化される。さらに、第２のイベントサンプル数

が、ロール方向変化からの時間間隔

と同様に、ゼロに初期化される。ここで使用する上添え字「Ｅ」は、エネルギアルゴリズム３００．２に関連する変数を示している。 In step (310.2), when entering energy algorithm 300.2 for the first time following step (306.1), energy algorithm 300.2 is initialized. Number of event samples n ^E and angular position values

Is initialized, ie set to zero. Also the sample time just before the event entrance time

Is the energy event entry time initialized to the value of the current time t

Time interval since initialization into the algorithm

Is initialized to a zero value. In addition, the number of second event samples

Is the time interval from roll direction change

Similar to, it is initialized to zero. The superscript “E” used here indicates a variable related to the energy algorithm 300.2.

が現行時刻ｔに等しく設定され、かつエネルギイベント時間

がエネルギイベント入口時刻

から現行時刻

までの時間間隔として次式で計算される。

ステップ（３２２．２）における、エネルギアルゴリズム３００．２の１つの出口基準は、例えば、エネルギイベント時間

が、最大時間間隔閾値

より大きいこと、すなわち次式である。

エネルギアルゴリズム３００．２の別の出口基準は、例えばエネルギイベント時間

が、最小時間間隔閾値

よりも大きく、かつステップ（３０６．２）からの入口基準が最近に満たされてからの時間が、第２の時間間隔閾値

より大きいこと、すなわち次式である。

During subsequent iterations of the energy algorithm 300.2, if ONGOING_ENERGY_EVENT_FLAG is set in step (304.2), the number of event samples n ^E is incremented in step (312.2) and the associated current Sample time

Is set equal to the current time t and the energy event time

Is the energy event entrance time

To current time

Is calculated as the time interval until

One exit criterion of the energy algorithm 300.2 in step (322.2) is, for example, energy event time

Is the maximum time interval threshold

It is larger, that is:

Another exit criterion for energy algorithm 300.2 is, for example, energy event time.

Is the minimum time interval threshold

Greater than and the time since the entry criterion from step (306.2) was recently met is the second time interval threshold.

It is larger, that is:

を約１２秒に設定し、最小時間間隔閾値

を約４秒に設定し、第２時間間隔閾値

は約２秒に設定した。したがって、この例については、エネルギアルゴリズム３００．２は、少なくとも４秒間、しかし１２秒を超えない時間において実行され、このような制約を受けて、最近に満足された入口基準からの時間間隔が２秒を超えると、外に出ることになる。エネルギアルゴリズム３００．１の外に出ると、ONGOING_ENERGY_EVENT_FLAGは、ステップ（３２０．２）でリセットされて、その後にステップ（３０６．２）で入口基準が満たされると、エネルギアルゴリズム（３００．２）に関連する変数がステップ（３１０．２）で初期化される。 The energy algorithm 300.2 is substantially prior to restarting (i.e., out of the algorithm and reset) than the measure algorithm 300.1 because a relatively slow rollover event may occur. Need a long time. In the case of certain types of vehicles, the time interval threshold is based on actual rollover data.

Is set to about 12 seconds and the minimum time interval threshold

Is set to about 4 seconds and the second time interval threshold

Was set to about 2 seconds. Thus, for this example, the energy algorithm 300.2 is run at a time of at least 4 seconds, but not more than 12 seconds, and with such constraints, the time interval from the recently satisfied entry criteria is 2 If you exceed 2 seconds, you will go out. Once outside of the energy algorithm 300.1, the ONGOING_ENERGY_EVENT_FLAG is reset in step (320.2) and is then associated with the energy algorithm (300.2) if the entry criteria are met in step (306.2). The variable to be initialized is initialized in step (310.2).

が、

の符号付の値を積分することによって、以下のように評価される。

ここで、積分時間増分ｄｔは、現行反復における時刻

と前回反復における時刻

の差として求められ、この差はサンプリング速度が均一の場合には一定であり、以下にようになる。

そして

は次式で求められる。

In step (322.2), if the exit criteria are not met, the algorithm calculation is updated in step (326.2) as follows for a particular iteration of energy algorithm 300.2:
First, the angular position

But,

Is integrated as follows to evaluate as follows.

Where the integration time increment dt is the time in the current iteration

And the time in the previous iteration

This difference is constant when the sampling rate is uniform, and is as follows.

And

Is obtained by the following equation.

において適切に補償されない可能性がある。エネルギアルゴリズム３００．２においては、アルゴリズム入口基準が満たされた最近の時刻の後、少なくとも

秒間、例えば２秒間は外に出ることはなく、これによってエネルギアルゴリズム３００．２の持続時間を

秒、例えば１２秒まで延長させ、このために角速度センサ２０からの信号における比較的小さなオフセット、例えば２〜３度／秒に対して、大幅なロール角積分誤差（例えば２４から３６度）につながる可能性がある。悪路においては、車両１２は実質的に振動ロール運動を示すことがあり、したがって［悪路イベント］では、角速度ω_ｘは、真の角速度オフセット

の回りに振動するという特徴がある。 In step (326.2), the algorithm calculation is adapted to compensate for the offset of the angular velocity ω _x signal, either due to a gyroscope error or an offset resulting from significant vehicle motion, otherwise , Especially for rough road conditions where the angular velocity may exhibit vibration behavior, these offsets are

May not be properly compensated for. In the energy algorithm 300.2, at least after the most recent time when the algorithm entry criteria is met,

Never go out for a second, for example 2 seconds, which increases the duration of the energy algorithm 300.2

Extending to a second, for example 12 seconds, this leads to a significant roll angle integration error (for example 24 to 36 degrees) for a relatively small offset in the signal from the angular velocity sensor 20, for example 2-3 degrees / second there is a possibility. On rough roads, the vehicle 12 may exhibit substantially oscillating roll motion, so in [bad road events], the angular velocity ω _x is a true angular velocity offset.

It has the feature that it vibrates around.

が−６．５度／秒である、角速度ω_ｘ信号を時間の関数としてプロットしてある。通常のロール事象においては、ロール事象中に

の符号は変化をしないために、

信号から悪路状態を認識することが可能である。これらの状況下で、積分されたロール角

は、

が符号を変えるたびに、ゼロに向かって以下の式に従って減衰される。

ここで、カウンタ

は、逆転時にイベントサンプル数

に等しく設定され、このイベントサンプルカウンタが、

が方向を変える度に、最近の方向変更からの時間間隔に応じて、０．１％から５０％の間の量でロール角

を減衰させる。 For example, referring to FIG. 13, the true angular velocity offset

The angular velocity ω _x signal is plotted as a function of time, with −6.5 degrees / second. In a normal roll event, during the roll event

In order not to change the sign of

It is possible to recognize a rough road condition from the signal. Under these circumstances, the integrated roll angle

Is

Each time changes sign, it is attenuated towards zero according to the following equation:

Where counter

Is the number of event samples when reversing

And this event sample counter is set to

Roll angle in an amount between 0.1% and 50%, depending on the time interval since the last change of direction each time the direction changes

Is attenuated.

の−６．５度／秒を、積分に先立ち除去した。第２の条件として、ロール角

は、偏りのある角速度ω_ｘデータから積分して、次いで前述したようにロール振動の補償をする。第３の条件としては、ロール角

は、上述のロール振動に対する補償なしで角速度ω_ｘデータから積分して求め、これは比較的長い積分区間について補償をしない角速度ω_ｘの偏りがある結果として、ロール事象の誤検出が発生する可能性を示すものである。上記のロール振動に対する補償は、角速度が実質的に一方向である実際のロール事象の検出に悪影響を与えることなく、ロール振動により導入された積分誤差を実質的に補正する。 Referring to FIG. 14, the effect of the above compensation on the roll vibration effect is shown, where the roll angle integrated from the angular velocity data plotted in FIG. 13 is shown as a function of time for various conditions. The first condition is true angular velocity offset

Of −6.5 degrees / second was removed prior to integration. As the second condition, the roll angle

Is integrated from the biased angular velocity ω _x data and then compensates for roll vibration as described above. As the third condition, the roll angle

Is obtained by integrating from the angular velocity ω _x data without compensation for the above-described roll vibration, which may result in false detection of roll events as a result of the bias of angular velocity ω _x that is not compensated for a relatively long integration interval. It shows sex. The compensation for roll vibration described above substantially corrects the integration error introduced by roll vibration without adversely affecting the detection of actual roll events where the angular velocity is substantially unidirectional.

ステップ（３２２．２）のアルゴリズム計算に続いて、ステップ（３３０．２）で評価されるアルゴリズム検出基準は、例えば図８ｃに示すような複数の検出条件を含む。すべての検出条件が満足される場合には、通常エネルギイベント閾値を超えて、ロールオーバが発生すると考えられ、ステップ（３４０）において安全化アルゴリズム２００からの、関連する安全化基準を満たせば、ステップ（３５０）において、関連する１人または複数の乗員への傷害を軽減するために、関連する１つまたは複数の安全拘束アクチュエータ３０が展開される。エネルギアルゴリズム３００．２の検出基準は、測度アルゴリズム３００．１について上述したものと類似の検出戦略に従って決定される。 In step (326.2), the calculation of the algorithm further gives a record of the most recent time that the entry criteria of step (306.2) are met, thereby providing an auxiliary basis for the exit criteria of step (322.2). Is given by

Subsequent to the algorithm calculation in step (322.2), the algorithm detection criterion evaluated in step (330.2) includes a plurality of detection conditions as shown in FIG. 8c, for example. If all detection conditions are satisfied, it is considered that a rollover will occur, usually exceeding the energy event threshold, and if the relevant safety criteria from the safety algorithm 200 in step (340) are met, the step At (350), the associated one or more safety restraint actuators 30 are deployed to mitigate injury to the associated passenger or passengers. The detection criteria for the energy algorithm 300.2 are determined according to a detection strategy similar to that described above for the measure algorithm 300.1.

およびロール角

の挙動と、それに関連する運動軌跡とに基づいている。ω‐θ位相空間の例を図１５に示してある。
剛体運動力学に従って、関連する剛体のロール事象と非ロール事象を判別する位相空間内の理論的閾値境界が存在する。例えば、この理論的閾値境界は次式で求められる。

ここで、ｍｇは車両の重量、Ｔは車両のトラック幅、Ｉは車両のロールの慣性モーメント、ｈ_ＣＧは車両の重心の高さである。この式は問題とする範囲ではω‐θ面内でほぼ線形である。 The main detection criterion of the energy algorithm 300.2 is in the phase space related to angular velocity and roll angle (ie, ω-θ phase space).

And roll angle

Based on the behavior of the robot and the motion trajectory associated therewith. An example of the ω-θ phase space is shown in FIG.
According to rigid body kinematics, there is a theoretical threshold boundary in the phase space that discriminates between related rigid body roll and non-roll events. For example, this theoretical threshold boundary is obtained by the following equation.

Here, mg is the weight of the vehicle, T is the track width of the vehicle, I is the moment of inertia of the roll of the vehicle, and h _CG is the height of the center of gravity of the vehicle. This equation is almost linear in the ω-θ plane in the range of interest.

However, because of the non-rigid effect, it is advantageous to model the actual threshold boundary as a piecewise linear boundary, usually consisting of about five or six connected segments along the theoretical threshold boundary above. May be adapted for a particular vehicle 12 or vehicle platform to improve the discrimination performance of roll and non-roll events. In general, this boundary is expressed as either a function in phase space (eg, a function of roll angle θ), a piecewise function in phase space (eg, a piecewise function of roll angle θ), or a table lookup in phase space. Can do.
Referring to FIG. 15, the actual rollover test data filtered using the above-described moving average filter is associated with the relevant theoretical threshold values for tests A and B shown in FIGS. 11a and 11b, respectively, according to the conditions of FIG. Plotted in the ω-θ phase space along with an example boundary and an example of an actual piecewise linear threshold boundary.

と実際の閾値境界の線分との距離が、その関連する端点の角度値

が現行ロール角

の上下限を決めている線分それぞれについて、各反復毎に計算される。実際の閾値境界の各線分は、その端点

によって定義される。現行順序組と実際の閾値境界の適切な線分との距離Ｄは次式で求められる。

ここで、この距離がゼロ未満の場合には、実際の閾値境界を超えたたことになる。 Current sequence pair

Is the angle value of the associated end point.

Is the current roll angle

For each line segment that determines the upper and lower bounds, it is calculated at each iteration. Each line segment of the actual threshold boundary is its end point

Defined by A distance D between the current ordered set and an appropriate line segment of the actual threshold boundary is obtained by the following equation.

Here, when the distance is less than zero, the actual threshold boundary is exceeded.

の軌跡の勾配は次式で求められる。

またω‐θ位相空間における、この勾配の関連する角度は次式で求められる。

ステップ（３３０．２）において、角度βが限度内にある（すなわち、β^min＜β＜β^max、ここで例えばβ^min＝７５度、β^max＝９０度）場合には、ロール角速度の絶対値は、時間と共に増加しており

実際の閾値境界までの距離が、ゼロより小さく

かつロール角θ^Eがロール角閾値θ^Thrよりも大きい

ここで例えばθ^Thr＝１０度）場合に、エネルギ検出基準が満足される。あるいは、ω‐θ位相空間における距離が、閾値D^Thrより小さく

かつロール角θ^Eがロール角閾値θ^Thrよりも大きい

場合に、エネルギ検出基準が満たされる。
エネルギ検出基準が、ステップ（３３０．２）で満足され、かつステップ（３４０）において、安全化基準が満たされる場合には、ステップ（３５０）において、関連する１人または複数の乗員への傷害が軽減するために、関連する１つまたは複数の安全拘束アクチュエータ３０が展開される。エネルギアルゴリズム３００．２の検出基準が満たされるときまでに安全化基準が満たされない場合には、安全化基準が満たされるか、またはステップ（３２２．２）でエネルギアルゴリズム３００．２から出るかのいずれかになるまで、エネルギアルゴリズム３００．２が反復されるように、エネルギアルゴリズム３００．２の展開決定はラッチされない。 in ω-θ phase space

The slope of the trajectory is obtained by the following equation.

The angle related to this gradient in the ω-θ phase space can be obtained by the following equation.

In step (330.2), if the angle β is within limits (ie, β ^min <β <β ^max , where, for example, β ^min = 75 degrees, β ^max = 90 degrees), the absolute value of the roll angular velocity Is increasing over time

The distance to the actual threshold boundary is less than zero

And roll angle θ ^E is larger than roll angle threshold θ ^Thr

Here, for example, θ ^Thr = 10 degrees), the energy detection criterion is satisfied. Alternatively, the distance in the ω-θ phase space is smaller than the threshold value D ^Thr

And roll angle θ ^E is larger than roll angle threshold θ ^Thr

If the energy detection criteria are met.
If the energy detection criteria are satisfied in step (330.2) and if the safety criteria are satisfied in step (340), then in step (350) there is an injury to the relevant passenger or passengers. To mitigate, the associated one or more safety restraint actuators 30 are deployed. If the safety criteria are not met by the time the energy algorithm 300.2 detection criteria are met, either the safety criteria are met or the energy algorithm 300.2 is exited in step (322.2). Until then, the deployment decision of the energy algorithm 300.2 is not latched so that the energy algorithm 300.2 is repeated.

It goes without saying that the measure algorithm 300.1 and the energy algorithm 300.2 can be executed in series or in parallel on the common processor 26 or on separate processors 26. If executed in series, the steps for one iteration shown in FIG. 6 are completed for one algorithm, and then the other algorithm is either from step (302) for the first pass or subsequent Step (150) is started for this path.
Although rollover detection algorithms have been represented by specific types of equations, it goes without saying that these calculations can be implemented on a particular processor 26 in a variety of ways without departing from the scope of the present teachings. Nor. For example, to actually implement a particular calculation described herein on a particular processor, for example, the resolution of the associated analog / digital converter, the type and accuracy of mathematical processing that can be performed by the particular processor 26, Further, there is a possibility that correction is required according to the word length of the specific processor 26 or the like.

を積分することによって求めるロール角の測度を利用すると説明したが、例えば傾斜センサを用いて測定したロール角を、計算によるロール角の代わりに使用することも可能である。 Although it has been described that the roll discrimination algorithm is applied to sampled data in the present application, it is needless to say that the algorithm can be continuously implemented using, for example, an analog processor. Furthermore, the number of event samples n ^M can be either explicit or implicit in the actual implementation of role discrimination, and the associated time-dependent variables are functions of time t, or event samples n ^M , n ^It should also be understood that any of the functions of ^E can be expressed.
Measure algorithm 300.1 and energy algorithm 300.2 are related

Although it has been described that a roll angle measure obtained by integrating is used, for example, a roll angle measured using an inclination sensor can be used instead of a calculated roll angle.

  With reference to FIGS. 16 and 17, a rollover detection system 10.1 according to another embodiment includes the rollover detection system 10 shown in FIG. 2 and described above, and further includes an occupant detection system 42, which occupant detection system. The proximity of the occupant 44 to one or more interior sides of the vehicle incorporating one or more safety restraint actuators 30 provided to mitigate associated rollover-induced injury to the occupant 44 It is adapted to detect the degree.

  The occupant detection system 42 at least detects the presence of the occupant 44 at a given seating location and measures the distance that the occupant is away from the inner boundary 46, and / or the occupant 44 from the vertical, eg, the inner boundary By measuring the angle of inclination towards 46, the degree of proximity or position of the occupant 44 or part thereof, for example the head of the occupant 44, relative to the internal boundary 46 of the associated vehicle 16 It can be further adapted to provide, directly or indirectly, with a measure for the location of the vehicle seat 48 to be. For example, the occupant detection system 42 can provide a measure of the lateral position or distance of the occupant head 50 or the torso 52 with respect to the associated internal boundary 46 of the vehicle 12, which is related to the associated head. It is useful for estimating the contact time of the part 50 or the torso 52, that is, the time that the occupant head 50 or the torso 52 is expected to contact the inner boundary 46 of the vehicle 12.

  The occupant detection system 42 includes various detection technologies such as visual sensing, capacitive sensing, active or passive infrared sensing, laser sensing, optical sensing, wireless (radar) sensing, microwave sensing, and sound waves (ultrasound). ) Sensing, multi-point seat weight distribution sensing, or electrostatic sensing can be incorporated. For example, in a capacitive sensing system (or electric field sensing system), the associated electrode (s) are on the roof rail directly above the interior roof or window, or on the pillar (or Other fixed positions can also be arranged. Alternatively, or in addition, electrodes can be placed at the headrest position or upper seat back of the associated vehicle seat 48. As another example, occupant 44, for example, using a vision system that incorporates an associated structured light illumination system and an associated camera and image processing system for detecting objects from interaction with structured light. Can also be detected.

ここで、Ｄ（ｔ）は、関連する内部境界４６からの、頭部５０または胴部５２の距離であり、Ｖ_{ｌａｔｅｒａｌ}およびＡ_{ｌａｔｅｒａｌ}はそれぞれ、例えば、関連する距離、速度および加速度の最新の推定値または計測値を使用する、頭部５０または胴部５２の関連する速度および加速度であり、ここで加速度は一定であると仮定している。 For an occupant detection system 42 that provides a measure of the distance between the occupant's head 50 or torso 52 and the associated internal boundary 46 of the vehicle 12 as a function of time, the associated distance over time. Can be tracked and the associated contact time can be estimated from them. For example, the associated lateral velocity of the head 50 or the torso 52 is estimated from the associated position change with respect to time or estimated directly by measuring this velocity, for example from the Doppler shift of the measurement signal. can do. The associated lateral acceleration of the head 50 or torso 52 can then be estimated from the change in associated velocity over time. For example, the associated speed can be estimated or measured continuously, or estimated or measured at discrete intervals. The associated contact time can then be estimated by solving the following equation for time:

Where D (t) is the distance of the head 50 or torso 52 from the associated internal boundary 46, and V _lateral and A _lateral are, for example, current estimates of the associated distance, velocity and acceleration, respectively. The associated velocity and acceleration of the head 50 or torso 52 using the value or measurement, where the acceleration is assumed to be constant.

The rollover detection system 10.1 incorporates a controller 54, for example as shown in FIG. 17, which is the lateral accelerometer 18 described in more detail above according to the embodiment shown in FIG. , Angular velocity sensor 20, filters 22, 24, processor 26, and memory 28.
A set of left 56.1 and right 56.2 occupant sensors can operate on one or more input ports of the processor 26, eg, via a single bus 48, multiple signal lines, or multiple signal cables It is connected to the. For example, FIG. 17 shows both left 60.1 and right 60.2 first occupant sensors, eg, visual sensors placed overhead, and left 62.1 and right 61.2 second occupant sensors, eg , Shows an embodiment including capacitive sensors installed in the associated left 64.1 and right 64.2 doors. Each occupant seating location is shown as having two different occupant sensors associated with it, but the particular number of occupant sensors is not limiting. For example, in another embodiment, each seating location to be detected may have only one occupant sensor 56.1, 56.2, or more than two in relation to any one seating location. There may be an occupant sensor.

  The processor 26 also includes a related set of left 30.1 and right 30.2 safety restraint actuators, eg, left 32.1 and right 32, eg, via signal bus 66, multiple signal lines, or multiple signal cables. .2 seat belt pretensioner, left 36.1 and right 36.2 trunk airbag inflator, left 38.1 and right 38.2 roll curtain, left 40.1 and right 40.2 overhead airbag Connected to the inflator. The processor 26 determines the direction of the roll from the associated signals of the angular velocity sensor 20 and / or the lateral accelerometer 18 so that one or more of the sides L, R on which the vehicle 12 is rolling are located. The safety restraint actuator 30 can be deployed.

  In general, it is desirable that the safety restraint actuator 30 be deployed only when needed, which is sufficient energy to avoid unnecessary associated repair costs and to protect the occupant 44 in an accident situation. This is to avoid an adverse effect on the occupant 44 located in the “danger” zone of the safety restraint actuator 30 that needs to have the Thus, the safety restraint actuator 30 mitigates injury to the occupant caused, for example, by head contact with the interior of the vehicle (e.g., window, upper door rail, ceiling), or by release or partial release from the vehicle. It would be beneficial to deploy in situations where the benefits of doing so exceed the risk of injury due to deployment of the safety restraint actuator 30. For example, the risk of injury from an airbag inflator generally increases as the occupant distance from the airbag inflator decreases. It is generally beneficial if the associated safety restraint actuator 30 can be fully deployed before it interacts with the occupant 44. Different safety restraint actuators 30 may require different amounts of deployment time. For example, an airbag inflator requires about 20 milliseconds to inflate, depending on the size of the associated airbag.

  As shown in the data of FIG. 10, rollover events are generally much slower than frontal or side impacts, so that a roll event is detected and the associated safety restraint actuator (s) 30 are deployed. It will give you quite a long time. For example, in test A, head contact does not occur until 368 milliseconds after energy algorithm 300.2 first detects a roll event and 59 milliseconds after safety algorithm 200 confirms a roll event. It was. In Test D, head contact did not occur until 98 milliseconds after the measure algorithm 300.1 detected a roll event. In both of these cases, there is sufficient time to deploy the associated safety restraint actuator (s) 30 long before the associated head contact, thereby causing the associated restraint actuator 30 (s) The probability of fully deploying the plurality) before their associated interaction with the occupant 44 can be improved. However, in certain situations, particularly tripped roll events, such as curb trip type roll events, detect whether a complete vehicle rollover occurs before the associated head contact or partial release. It is impossible. This can vary, for example, depending on where the occupant is seated relative to the internal boundary 46 of the vehicle 12 with which it is associated. In these situations, the associated safety restraint actuator (s) 30 are potentially activated in response to an event that the vehicle will not eventually roll over so that they are deployed prior to head contact. Where it is desirable to avoid the adverse effects that result from deploying the safety restraint actuator (s) 30 after head contact by activating before risking and would otherwise be activated There is.

  According to one mode of operation, in response to detection of a roll event by the measure algorithm 300.1 or energy algorithm 300.2, as confirmed by the safety algorithm 200, the occupant detection system 42 may In the absence of an occupant on the right side, it is possible to prohibit activation of the associated left 30.1 or right 30.2 safety restraint actuator (s). For example, if there is no occupant on the right side of the vehicle 12 and the vehicle 12 is about to roll to the right, the left 56.1 occupant sensor (single or In response to input from the plurality, the processor 26 does not activate the left 30.1 safety restraint actuator (s) in response to the roll, thereby unnecessarily left 30.1 safety restraint actuator ( Potential adverse effects of deploying the singular (s), such as repair costs, can be avoided.

  According to another mode of operation, the timing of activation of the safety restraint actuator (s) 30 is responsive to the distance of the occupant 44 from the associated internal boundary 46 of the vehicle 12 or to a detected occupant motion state. It can be adapted to respond to expected head or torso contact times that are responsive. For example, in response to a measure of occupant position from occupant detection system 42, or prediction of occupant position therefrom, activation of safety restraint system (s) 30 is prohibited or otherwise rolled event Can be delayed or advanced with respect to the time at which is detected.

  For example, until head / torso contact with the inner boundary 46 of the associated vehicle 12 is imminent [depending on how long the associated safety restraint actuator (s) need to be deployed Up to as much as 10-30 milliseconds before; ie, if the head or torso contact time is greater than the deployment time of the associated safety restraint actuator (s) 30], some or all of the safety restraint actuators Provide the rollover detection system 10.1 with as long a time as possible to delay the deployment of the 30 (s) to determine or best predict whether the event will lead to a complete vehicle rollover. You can also. By making deployment decisions based on occupant position or head contact, the maximum amount of time to estimate the associated acceleration and angular velocity measures is used to accurately distinguish between roll and non-roll events. Provided to the detection system 10.1.

  As another example, if imminent head contact is detected or expected by occupant detection system 42 [eg, head / torso contact time is deployed in associated safety restraint actuator (s) 30 If less than or equal to time], the deployment of some or all of the safety constraint actuator (s) 30 may be accelerated, the measures algorithm 300.1, energy algorithm 300.2, safety algorithm 200 related It can also be done by reducing one or more thresholds to a low but still large level. For example, if imminent head contact is detected or expected by occupant detection system 42, if a set of minimum conditions is met, for example, angular velocity is associated with a minimum threshold, For example, some or all safety restraint actuators (if greater than about 60 degrees / second and / or if the lateral acceleration exceeds an associated minimum threshold (eg, at least 1 g for at least 10 milliseconds) ( (Single or plural) 30 can be deployed.

  As another example, when impending head contact is detected or expected by occupant detection system 42, for the measure algorithm 300.1 and energy algorithm 300.2, respectively, shown in FIGS. One or both of the associated threshold functions can also be reduced compared to a system without the occupant detection system 42. These thresholds can be further adjusted according to the occupant head contact measurement speed so that a lower threshold is provided according to the fast lateral head speed. Any imminent device that can be reset or reused is likely to be deployed after imminent head contact is recognized and these small thresholds are exceeded. As an alternative, all rollover restraints, including airbags, can be deployed in time to prevent the initial collision of the occupant head with the associated internal boundary 46 of the vehicle 12. The occupant lateral movement provides another indication that a potential rollover event is in progress and can be relied upon as a rollover safety measure independent of other sensing signals such as roll speed.

  As another example, if the decision to deploy the safety restraint actuator (s) is not made until head or torso contact is detected by the occupant detection system 42, the safety restraint actuator (s) 30 can be selectively deployed depending on the position of the occupant. For example, a seat belt pretensioner 32.1, 32.2 or overhead airbag inflator 40.1, 40.2 is relatively harmless to an occupant located against the associated internal boundary 46 of the vehicle 12. Therefore, it is also possible to deploy these immediately. However, the deployment decision of roll curtain 38.1, 38.2 or torso airbag inflator 36.1, 36.2 will properly deploy the bag between occupant 44 and associated internal boundary 46 of vehicle 12. Delay until the opportunity is given, thereby preventing the occupant 44 from being positioned between the safety restraint actuator 30 and the vehicle door 64.1, 64.2 or side window 39.1, 39.2. At the same time, the protection of the passenger 44 can be strengthened.

  The roll identification algorithm described here is derived from the relevant test data and may therefore require adjustment when applied to other types of vehicles that are different from the vehicle from which the parameters were extracted. The criteria for this adjustment are, for example, robust and early detection of rollover events and at the same time, avoid misidentifying non-rollover events as rollover events to the extent possible. That is. The specific values for the various parameters described in this application are not limited thereto and may vary for different types of vehicles, for example, where the likelihood of rollover is different. For example, a vehicle with a relatively high center of gravity or a vehicle with a relatively narrow wheelbase, such as a sport utility vehicle, is more likely to generate a rollover than a vehicle with a relatively low center of gravity or a wide wheelbase, such as a riding sedan. In addition, the rollover detection system 10 can be adapted to detect pitch over events, i.e. events around the local Y axis of the vehicle, by including an associated longitudinal accelerometer and pitch velocity sensor.

The occupant sensor enables an improved rollover detection system by allowing modification of the deployment of the safety restraint system in response to either occupant presence, position, estimated contact time or contact speed, thereby rolling Reduces the possibility of occupant injury due to an over event or deployment of an associated safety restraint system.
While specific embodiments have been described in detail, those skilled in the art will be able to develop various modifications and alternative approaches to these details with reference to the general teachings of the present disclosure. You will understand. Accordingly, the specific arrangements disclosed are for purposes of illustration only and are not intended to limit the scope of the invention, which is intended to cover the full scope of the appended claims and all equivalents thereof. Should be awarded.

It is a rear view of the vehicle before roll generation start. It is a rear view of the vehicle in which the roll is generated. It is a block diagram which shows a rollover detection system. It is a flowchart which shows a rollover detection algorithm. FIG. 5 is a flow diagram illustrating a data collection and processing algorithm incorporated into a rollover detection algorithm. It is a flowchart which shows the safety algorithm incorporated in the rollover detection algorithm. It is a flowchart which shows a rollover detection algorithm. FIG. 5 is a flow diagram illustrating an algorithm for determining whether sensor recalibration is required, incorporated in a rollover detection algorithm. It is a table | surface which shows the detail of a rollover detection algorithm. It is a table | surface which shows the detail of a rollover detection algorithm. It is a table | surface which shows the detail of a rollover detection algorithm. It is a table | surface which shows the example of the parameter value of a rollover detection algorithm. It is a table | surface which shows the example of the parameter value of a rollover detection algorithm. FIG. 5 is a table showing conditions associated with various rollover and non-rollover events. FIG. FIG. 4 is a graph of filtered roll angular velocity, roll angle, and filtered lateral acceleration for a vehicle that has undergone a corkskrew rollover test, designated Test A, in which a rollover event occurs. FIG. 7 is a graph of filtered roll angular velocity, roll angle, and filtered lateral acceleration for a vehicle that has undergone a spiral rollover test, designated Test B, in which no rollover event occurs. FIG. 6 is a graph of filtered roll angular velocity, roll angle, and filtered lateral acceleration for a vehicle that has undergone a deceleration sled test named Test C with no rollover event occurring. FIG. 6 is a graph of filtered roll angular velocity, roll angle, and filtered lateral acceleration for a vehicle that has undergone a deceleration sled test named Test D, in which a rollover event occurs. FIG. 6 is a graph showing the rollover measure merit index and the associated deployment threshold as a function of time for a test D rollover event and a test C non-rollover event, depending on the measure algorithm. FIG. 6 is a graph showing roll angular velocity as a function of time for a signal including roll angular velocity offset. FIG. 14 is a graph showing roll angle as a function of time based on the data of FIG. 13 for various related processes for identifying roll angle from roll angular velocity. FIG. 6 is a graph showing roll angular velocity as a function of roll angle and associated rollover threshold according to an energy algorithm for a test A rollover event and a test B non-rollover event. It is the front view which shows the compartment of a vehicle, and the block diagram which shows another embodiment of a rollover detection system. It is a block diagram of the controller by the rollover detection system shown in FIG.

Claims (23)

a. A roll angular velocity sensor operably coupled to a vehicle and adapted to measure a roll speed about a roll axis of the vehicle, wherein the roll axis substantially coincides with a longitudinal axis of the vehicle Roll angular velocity sensor;
b. An occupant sensor that provides a measure of proximity or position to an internal boundary of the vehicle in response to the presence or position of the occupant in the vehicle; and c. Operatively coupled to the roll angular velocity sensor and the occupant sensor to generate a signal for controlling a safety restraint system in response to detecting a predicted rollover condition responsive to a signal from the roll angular velocity sensor A processor adapted to include a processor adapted to modify a detection criterion associated with the signal for controlling the safety restraint system in response to a signal from the occupant sensor. A system that detects the rollover condition of a vehicle. The system for detecting a rollover condition of a vehicle according to claim 1, wherein the occupant sensor is responsive to the position of the occupant's head. The system for detecting a rollover condition of a vehicle according to claim 1, wherein the occupant sensor is responsive to the position of the torso of the occupant.   The occupant sensor is selected from capacitive sensors, active infrared sensors, passive infrared sensors, visual sensors, laser sensors, optical sensors, radar sensors, microwave sensors, acoustic wave sensors, multi-point seat weight distribution sensors, and electric field sensors The system for detecting a rollover condition of a vehicle according to claim 1, comprising at least one sensor. A processor is adapted to determine a roll angle measure by integrating a signal from the roll angular velocity sensor, and a signal for controlling a safety restraint system is further responsive to the roll angle measure; The system for detecting a rollover state of the vehicle according to claim 1. An accelerometer operably coupled to the vehicle and adapted to measure substantially lateral vehicle acceleration, and a signal for controlling the safety restraint system is a signal from the accelerometer; The system for detecting a rollover condition of a vehicle according to claim 1, further responsive to . A method for enabling detection of a rollover condition of a vehicle,
a. Enabling the acquisition of a measure of the roll angular velocity of the vehicle about a roll axis substantially coincident with the longitudinal axis of the vehicle;
b. Enabling obtaining a measure of occupant position in the vehicle; and c. Enabling the generation of a signal for controlling the safety restraint system, wherein the signal for controlling the safety restraint system is responsive to a detection criterion along with a measure of the roll angular velocity, wherein the detection criterion is Responsive to a measure of the occupant position. 8. The method of enabling detection of a vehicle rollover condition according to claim 7, wherein a deployment threshold associated with the detection criteria is varied by an offset responsive to a measure of occupant position. 8. The method of enabling detection of a vehicle rollover condition according to claim 7, wherein the deployment threshold associated with the detection criterion is scaled by a factor responsive to a measure of occupant position. The method of enabling detection of a vehicle rollover condition according to claim 7, wherein a deployment threshold associated with the detection criterion is responsive to a rate of change of the occupant position measure. a. Enabling identification or acquisition of a roll angle measure from a roll angular velocity measure;
b. Enabling the identification of a threshold function in the phase space between the roll angular velocity measure and the roll angle measure;
c. Enabling modification of the threshold function in response to a measure of occupant position; and d. Further comprising enabling comparison between a measure in phase space that includes a combination of the measure of roll angular velocity and the measure of roll angle and the threshold function, the signal for controlling a safety restraint system comprising: The method of enabling detection of a vehicle rollover condition according to claim 7 responsive to an operation of comparing a measure in phase space with the threshold function.   The method of enabling detection of a vehicle rollover condition according to claim 11, wherein the threshold function comprises either a function in phase space, a piecewise function in phase space, or a table lookup in phase space.   The vehicle rollover condition of claim 11, wherein the operation of modifying the threshold function includes identifying a roll angular velocity offset in response to a occupant position measure and subtracting the roll angular velocity offset from the threshold function. To enable detection of The operations that enable the generation of signals to control the safety restraint system
a. Enabling the acquisition of a measure of the vehicle's lateral acceleration;
b. Allowing a particular figure of merit in response to a measure of the measure and the roll angular velocity of the lateral acceleration;
c. Enabling identification of a goodness index threshold in response to a measure of occupant position; and d. 8. The method of enabling detection of a rollover condition of a vehicle according to claim 7, comprising enabling detection of a rollover condition by comparing the goodness index and the goodness index threshold. a. estimating a contact time at which at least a part of the occupant contacts an internal part of the vehicle;
b. Comparing the contact time to a first threshold value corresponding to a time interval required to deploy the safety restraint system before interacting with the occupant; and c. 8. The detection of a vehicle rollover condition according to claim 7, further comprising adapting a signal for controlling the safety restraint system in response to an operation of comparing the contact time with the first threshold. How to make it possible.   The method of enabling detection of a vehicle rollover condition according to claim 15, wherein the operation of estimating the contact time includes estimating or measuring an occupant speed relative to the vehicle.   16. The vehicle rollover state of claim 15 further comprising storing measures of occupant position at different times, and the operation of estimating contact time includes estimating or measuring occupant acceleration relative to the vehicle. How to enable detection. The method of enabling detection of a vehicle rollover condition according to claim 15, wherein the operation of estimating the contact time is responsive to a plurality of measures of occupant position at different times.   16. The method of enabling detection of a vehicle rollover condition according to claim 15, further comprising modifying a signal for controlling the safety restraint system if the contact time is less than a first threshold.   20. The operation of modifying a signal for controlling the safety restraint system includes deploying the safety restraint system earlier than otherwise performed in response to detecting an expected rollover condition. A method for enabling detection of a rollover condition of a vehicle according to claim 1.   20. The method of enabling detection of a vehicle rollover condition according to claim 19, wherein the act of modifying a signal for controlling a safety restraint system comprises prohibiting deployment of the safety restraint system. The method of enabling detection of a vehicle rollover condition according to claim 15 , wherein the act of modifying a signal for controlling the safety restraint system comprises delaying deployment of the safety restraint system.   23. A vehicle roll according to claim 22, wherein the operation of delaying deployment of the safety restraint system includes delaying the deployment until the difference between the contact time and the first threshold is less than the second threshold. A method that allows detection of over conditions.

Images